The Cybersecurity and Infrastructure Security Agency finally got a new nominee for director, with President Donald Trump on Monday nominating Michael LoSapio, chief information security officer at Palantir.

CISA has been without a Senate-confirmed director for the entirety of Trump’s second term, after his first pick, Sean Plankey, languished in the Senate before withdrawing in April.

LoSapio could face some obstacles of his own, with Palantir far from a popular company with the Democrats set to take over Congress. Democrats have scrutinized the company over its work with the Department of Homeland Security, even as Palantir has denied aiding mass surveillance of immigrants.

As of late last month, Sen. Ron Wyden, D-Ore., was also still waiting for CISA to release a public version of a report on telecommunications cybersecurity, and he had placed one of several Senate holds on Plankey pending the receipt of that report.

Republicans will still control the Senate until January, however, and a lame-duck session began Monday.

Some CISA watchers were convinced that the agency’s acting director, Nick Andersen, was likely to continue holding the role. He was running out of days to hold the “acting” position under the law, but putting forward a nominee could pause the clock on that deadline.

The administration was thought to be looking for someone from a high-profile tech company to fill the CISA director position, a bill LoSapio fits.

The post Trump picks Palantir’s LoSapio to head CISA appeared first on CyberScoop.

I want to talk about AI hacking. There are three aspects: people hacking AI systems, people using AI to hack other systems, and AIs hacking systems on their own, not even maliciously. The last two matter more than the first. The third is the weirdest and potentially the most destabilizing.

I started writing about this in 2021 and in my 2022 book, A Hacker’s Mind. Back then I said it was mostly still science fiction, but not stupid science fiction. We’re seeing it more and more, especially in the past year. The trends I’m describing don’t require a major breakthrough in AI. They’re starting to happen now, and they’ll happen more often as AIs get better at autonomous problem-solving. And it will involve far more than computer code.

AI will hack all sorts of social, economic and political systems. It will find vulnerabilities and exploit them. This is a difference in kind, not just degree. We’re not ready for it, and we need to think about the implications now.

This article is adapted from a keynote speech delivered at the DEF CON 34 conference in Las Vegas on August 7, 2026.

First, I want to broaden the term “hacking.”

The tax code isn’t computer code, but it is code. It’s a series of algorithms with inputs and outputs. It has vulnerabilities, which we call loopholes, and exploits, which we call tax avoidance strategies. Black hat hackers exploit them for profit. We call them accountants and attorneys.

What is a hack? One definition: “Something that a system permits, but is unanticipated and unwanted by the designers.” It’s subjective and implies novelty and cleverness. A hack follows the rules of a system but subverts its goals or intent, just as a software hack follows the logic of code while doing something its authors neither anticipated nor wanted.

All systems of rules can be hacked, including professional sports, customer reward programs, financial systems and politics. Hacks even work against our cognitive functions. A curved hockey stick is a hack, and we know the name of the man who invented it. Mileage runs were a hack. The filibuster was originally a hack in Roman times. Gerrymandering is a hack. Hedge funds and private equity are full of them.

A system is just a set of rules or norms, and the rules aren’t always formal. Even the best-designed rules will be incomplete, inconsistent or ambiguous. As long as people want to subvert a system’s goals, there will be hacks.

AIs are becoming hackers. After decades of slow progress, they’ve gotten very good at finding and exploiting vulnerabilities in computer code. They do this largely autonomously, and they’re getting better scarily fast. But the implications extend far beyond software, to the tax code, financial regulations and all sorts of social, economic and political systems.

There are two distinct issues. First, someone might instruct an AI to hack one of these systems, feeding it the world’s tax codes or financial regulations to find profitable loopholes. Second, an AI might hack a system on its own, inadvertently. Both are dangerous, but the second is worse because we might never know it happened.

Now, reward hacking. AIs don’t solve problems the way people do. They stumble on solutions we never anticipated, and some subvert the intent of the system. They don’t share our context and don’t necessarily understand the implications, norms and values we take for granted.

Reward hacking is when an AI achieves a goal in a way its designers neither wanted nor intended.

Early pre-LLM examples are telling. In a two-person soccer simulation, an AI learned that kicking the ball out of bounds forced the opposing goalie to throw it in and leave the goal undefended. In a stacking simulation, an AI learned to flip a block upside down rather than stack it. In an evolution simulation where the goal was to cross a distant finish line, an AI made itself tall enough to fall over the line instead of growing longer legs or stronger muscles.

These are all hacks, the result of poorly specified goals or rewards. Today, they’d come from bad prompting.

The problem is universal. One researcher tried to teach a robot vacuum not to bump into things. Instead, it learned to drive backward, because there were no sensors in the back.

Any capable AI will find hacks wherever the rules have problems, inconsistencies or loopholes that still produce an acceptable solution. We see a hack because we understand context. The AI doesn’t, at least not in the same way.

This problem shows up often in myth.

In the story of King Midas, the god Dionysus grants him a wish: everything he touches turns to gold. Midas ends up starving when his food, drink and daughter all turn to gold. It’s a specification problem. He programmed the wrong goal.

In the Golem of Prague, a ghetto animates a clay statue to guard its residents. It guards them past all reason until they figure out how to turn it off. This is a guardrail problem. The villagers never limited the golem’s actions.

More generally, think of this as the genie problem. Genies are precise about the wording of wishes and can be maliciously pedantic when granting them. There’s no way to outsmart a genie. Whatever you wish for, he can grant it in a way you wish he hadn’t.

The reason is that in human language and thought, goals are always underspecified. We never include every caveat, exception and proviso, and we never close off every avenue for hacking. Any goal we specify will be incomplete. That’s largely fine in human interactions, because people understand context and usually act in good faith.

If I asked you to get me coffee, you’d pour a cup from the nearest coffee pot or walk to the corner shop. You wouldn’t bring me a pound of raw beans, buy a coffee plantation or snatch a cup from a stranger’s hands. I wouldn’t have to specify any of that. You’d just know.

Similarly, if I asked you to develop a technology that turns things to gold on touch, you wouldn’t build it to starve the person using it. You’d just know.

In 2015, Volkswagen was caught cheating on emissions tests. Human engineers programmed a regular computer to detect testing and behave differently during it. The cheat went undetected for years because it’s hard to tell what software is doing.

If I asked you to design engine control software that maximizes performance while passing emissions tests, you wouldn’t design it to cheat without realizing you were cheating. An AI would, because it will think “outside the box” when it has no concept of the box.

Unless programmers specify that the system must not behave differently during testing, an AI might come up with the same hack. The programmers would be satisfied, the accountants would be ecstatic, and no one would know until someone audited the code.

We can now constrain the goal to avoid that particular hack, but programmers won’t anticipate others. The lesson of the genie is that there will always be hacks programmers don’t anticipate.

We’re building AI genies: systems that take broad instructions and figure out how to carry them out, sometimes in ways we didn’t expect or intend. That’s the whole point of vibe coding. You underspecify what you want and let the AI work it out. Sometimes that produces genie behavior.

We can’t completely specify goals to an AI, and AIs can’t fully understand context. More precisely, current AIs have a shallow grasp of concepts. If we asked a current LLM to get me coffee, it would almost certainly do the right thing.

I worry about goals that aren’t well represented in training data, and about hacks that aren’t obvious, whose effects are subtle, like the Volkswagen hack.

Recommendation engines are an early example. They push people toward extreme content. No one programmed that. It emerged because that’s what people respond to. It didn’t take a bad actor. A basic automated system with feedback found the hack on its own.

We’re now seeing a second generation with today’s AI coding systems. AI researcher Simon Willison asked Fable to track down a stray scrollbar in a web app. When he returned, it had opened browsers, written its own screenshot tools, edited the framework’s internal templates to make the bug appear on cue and set up a local web server to collect measurements. It found the bug, but it also did many things he never asked for. Another researcher gave Fable a complex task, and it tried to hack its own memory limits because it needed more storage.

A new OpenAI model did something similar when it hacked into Hugging Face. It was being benchmarked on its ability to turn vulnerabilities into exploits. It decided the best way to reach its goal was to break out of its containment, access the internet and break into the Hugging Face network, because it thought the answers would be there.

Today’s LLMs are acting like genies, especially when powerful models are combined with sophisticated harnesses.

There are two kinds of genies. The Dionysus genie does the wrong thing. It understands your prompt literally but not correctly, and returns a mess you never intended. The Golem genie does the right thing but tramples everything in its path to get there. The paperclip maximizer is a Golem genie.

Consider some examples. Ask an AI to deal with spam calls, and a Dionysus genie might change your phone number. Ask it to book a particular flight, and a Golem genie might hack the airline database because the flight is full. Ask it to get a refund for an online purchase, and a Dionysus genie might draft a legal threat on fake letterhead and send it to the company. Ask it to buy tickets to a popular concert, and a Golem genie might discover the ticketing system is rate-limited and spin up thousands of cloud servers to hammer it. You’d know if some of these things happened, but you might never realize if others did. More detailed prompting won’t fix this. The lesson of the genie is that no amount of detail is good enough.

AI researchers aren’t surprised by any of this. Many are working on defenses against reward hacking and subtler forms of genie behavior. Some of this falls under “value alignment”: How do we teach AIs context? And how an AI satisfies a goal matters as much as whether it does. I want to build a benchmark to measure genie-like behavior, so we know where models stand and whether alignment is improving.

Think beyond LLMs and computer code. Consider specialized AIs built to find loopholes, and hacks of more general systems: the tax code, financial regulations and democratic procedures.

Financial systems are the most likely first target, because their rules are designed to be algorithmically tractable. High-frequency trading is itself a hack. Imagine giving an AI all of the world’s financial laws, plus all the news and financial data, and the goal of finding profitable loopholes. I suspect large investment banks are already doing this. The results will be novel tricks that are legal but unintended and unwanted, some of which might break the financial system.

Tax loopholes are the next likely target, for the same reasons. For a while, AIs will identify potential loopholes, and human hackers, namely attorneys and accountants, will decide which are feasible.

Further out, expect hacks of regulatory, legislative and procedural systems, and of the broader economy.

Hacking is as old as humanity. We’re creative problem solvers and loophole exploiters. We manipulate systems to serve our interests and strive for more influence, power and wealth.

Still, no human pursues their interests without constraint. Even sociopaths are limited by the complexity of society and their own contradictory impulses. They worry about their reputation and punishment, and they have limited time. These human qualities have limited hacking.

Hacking changed as everything became computerized. Computers are uniquely hackable because of their formalism and complexity. Today, nearly everything is a computer, from cars and appliances to phones. Our social, economic and political systems, including taxation, finance and elections, all involve computers.

Until recently, hacking was exclusively a human activity. Finding a novel hack required expertise, creativity, time and luck.

AI hacking will change everything again. AIs aren’t bound by the constraints that limit people. They’ll think like aliens, and they’ll change hacking’s speed, scale, scope and sophistication, sometimes at our direction and sometimes on their own.

Speed is the easiest to see. A creative process that takes a human months or years could be compressed into hours or seconds.

Scale is next. Once AI systems start discovering hacks, they can exploit them at a scale we’re not ready for. We already see this with AI-generated slop, which is rampant on social media. AI bots post around the clock, sometimes overwhelming real human discussion. It’s hard to know how much of social media is still people talking to people. What looks like boisterous political debate may be bots arguing with other bots, artificially shaping what we think and what we think others think.

Scope makes hacks more dangerous. AIs already make decisions that affect our lives, decisions we once thought belonged only to humans. As AI systems grow more capable, society will hand them more decisions, and more important ones. Hacks of these systems will do more damage.

Finally, AI hackers are more sophisticated. They can hold more variables in working memory and generate more complex hacks. What might happen if you fed an AI the entire U.S. tax code? Would it figure out, without being told, that it’s smart to incorporate in Delaware and register a ship in Panama? Consider the Double Dutch Irish Sandwich, the strategy Apple and Google used to avoid paying U.S. taxes. It involves U.S., Dutch and Irish law and an offshore tax haven. How many loopholes will an AI find that we don’t already know about? Dozens? Hundreds? Thousands? We have no idea.

These hacks will largely be perpetrated by the powerful against the rest of us. AI is fundamentally a power-enhancing technology, and so is hacking. The more power you already have, the more a hack gives you. If you or I found a tax hack, we might make a few thousand dollars. If a large investment bank found one, it and its clients could make billions.

Our societal systems for dealing with hacks were built when hackers were human, and they reflect the pace of human hacking. We have no system of governance that can handle hundreds, let alone thousands, of newly discovered tax loopholes. We won’t be able to recover from an AI finding unanticipated but legal hacks of financial systems, operating at AI speed, scale, scope and sophistication. Hacking becomes a problem society can no longer manage.

Now, defense. AIs that find software vulnerabilities help defenders as well as attackers. Vulnerabilities can be patched, removing them permanently, or turned into signatures that security systems use to detect and block attacks. In theory, AI could identify and patch every automatically discoverable vulnerability before a product ships. This is rapidly becoming routine in software development. We call it vulnops. We can imagine a future where accidental software vulnerabilities are a thing of the past.

Deliberate vulnerabilities are different. There will be an arms race. AI will craft increasingly obscure back doors, and AI will find them. It becomes a contest between the speed of creation and the speed of patching, and between obfuscation and reverse engineering.

The same logic applies to broader systems. Yes, AI hackers might find hundreds of vulnerabilities in the existing tax code. But the same technology can evaluate potential loopholes in any proposed tax law or ruling. A legislator, a watchdog group or a reporter could take the text of a bill and identify every exploitable loophole before it passes. The difference is that, unlike software, knowing about a tax loophole doesn’t mean it gets patched. In theory, though, we can fix them before the rich and powerful exploit them. So maybe defense will come out ahead here, too.

That requires resilient governing structures that can respond quickly. It won’t help if updating the tax code takes years, or if a legislative hack becomes so entrenched that it can’t be fixed politically. Maybe a lobbyist put it there on purpose. Maybe it was discovered after the law passed, and now lobbyists want to keep it. Microsoft can patch the vulnerabilities it finds. Patching is much harder when legislatures have to do it. We need society’s rules and laws to be as patchable as our computers.

This is a hard problem of modern governance, and it goes well beyond AI. It’s the problem of building governing structures for the speed and complexity of the information age.

The umbrella idea is trust. What do we need for AI systems to be truly trustworthy? Capability isn’t enough.

AIs won’t be trustworthy if the companies controlling them spy on us or subtly manipulate us. I call this the friend vs. tool problem. Will we have AI agents, or AI double agents?

AIs won’t be trustworthy if they’re insecure. How do we know they haven’t been manipulated? Almost all AIs operate in adversarial environments, so someone will benefit from hacking them. Unless we know they’re secure, we shouldn’t trust their output.

And AIs won’t be trustworthy if they act like genies. We need to know they’ll perform tasks the right way, not just perform the tasks we ask for.

I’ve written extensively about the need for “integrous” AI, including how to think about the problem, what a research agenda looks like and what good laws and policies look like. I predict integrity will be the key security problem of this decade, just as confidentiality was the key problem of the last one. This applies to both AI and the internet of things.

The big shift is agency, which I think of as Web 3.0. Computers now affect the physical world directly and with growing autonomy. Sensors go from data sources to attack surfaces, and actuators go from outputs to attack capabilities. An integrity failure no longer means bad spreadsheet data. It means the wrong dosage, the wrong valve position or the wrong braking distance.

Integrity isn’t one thing. I break it into input integrity, processing integrity, storage integrity and contextual integrity. Each is complex and important, and each deserves its own talk. I’ve written about them, and I hope you’ll read it.

Much of this isn’t really an AI problem. It’s not a technical problem so much as a capitalism problem or a democracy problem.

The failures of late-stage capitalism and democracy in the information age both stem from the rich and powerful getting too good at hacking our social, economic and political systems. This is the core argument of my 2021 book.

Seen this way, AI is exacerbating existing problems. Human inefficiency has largely kept those problems in check, and that won’t work much longer.

So consider this a call to action. Technically, we need to understand how AIs hack systems, build systems that are resilient to hacking and build AIs that won’t act as genies. Politically, we need to learn to govern and regulate at the speed of technology. Otherwise, market failures will overwhelm our ability to govern them.

I said hacking is as old as humanity and that AIs will overwhelm our ability to deal with it. The solutions aren’t specific to AI, but AI makes them more urgent. We need to understand them, and we need to implement them.

The post AI doesn’t have to go rogue to break the world appeared first on CyberScoop.

Three men concocted an elaborate, yearslong scheme that defrauded companies of at least $129 million by using AI to impersonate executives and lawyers, tricking targeted victims into transferring large payments to criminal-controlled accounts, the Justice Department said Thursday.

Jacob Soall, Sharon Hazan and Guy Sheetrit allegedly used AI to impersonate the voices and images of chief executive officers at companies and lawyers at major international law firms. The trio are accused of targeting thousands of victims between July 2024 and February 2026, according to court records.

Victims include a U.K.-based financial tech company that lost almost $11.6 million, a Florida-based pet food company that lost $5.2 million, a Missouri-based hospital system that lost $4.8 million, a Massachusetts-based mutual fund company that lost $4.6 million, a California-based hospital system that lost nearly $3 million, a New York-based investment company duped out of $2.8 million, and a Rhode Island-based bank that transferred nearly $2.7 million.

The trio’s scheme typically started by targeting accounting staff, pretending to be the CEO, and requesting help on a financial transaction, officials said. Using AI to create deepfakes pretending to be CEOs, the crew allegedly directed employees to contact attorneys who they would also impersonate to trick the employee into thinking the financial transaction was legitimate.

Prosecutors said the men and their co-conspirators identified and targeted employees who had sufficient authority to authorize large wire transfers. 

The scheme typically began with an email from an e-signature company, wherein the impersonated CEO claimed they were working on a “confidential financial transaction” that required the victim employee to sign a non-disclosure agreement before contacting the purported attorney at a victim law firm, according to an unsealed indictment.

Soall, Hazan, Sheetrit and their co-conspirators allegedly impersonated attorneys at some of the 50 largest law firms globally. 

“Super intelligence technology offers tremendous potential, but the public should recognize that there are some who seek to use this emerging technology to pursue criminal goals,” Jamie McDonald, U.S. attorney for the Southern District of New York, said in a statement. 

Prosecutors said the wire transfers were sent to bank accounts controlled by co-conspirators around the world, including the United States, Spain, Portugal, Greece, the United Arab Emirates, Singapore, Austria, Slovakia and China. The crew allegedly converted the funds to cryptocurrency and distributed the illicit proceeds amongst themselves and their co-conspirators.

Soall, a 41-year-old from Marbella, Spain, was arrested in Spain on June 17. He allegedly primarily impersonated attorneys from the United Kingdom, reading from a handwritten script designed to dupe targeted victims.

Hazan, a 54-year-old from Israel, was arrested in Spain on Aug. 5. He allegedly boasted that he had been laundering money since 1994, including more than $400 million in the past three years, and was previously convicted in the Southern District of New York of conspiracy to commit bank fraud in 2008.

Sheetrit, a 42-year-old from Israel and the United Arab Emirates, was arrested in the United Kingdom on Sept. 7 and allegedly laundered funds from the scheme.

The trio are charged with conspiracy to commit wire fraud, wire fraud, two counts of aggravated identity theft, money laundering conspiracy and two counts of money laundering. The seven counts carry a maximum sentence of 104 years in prison.

“The alleged scheme orchestrated by Soall, Hazan, and Sheetrit demonstrates how the use of advanced technology can be exploited to commit large scale fraud,” James C. Barnacle Jr., assistant director in charge of the FBI New York field office, said in a statement.

“As criminals continue to weaponize emerging technologies, the FBI remains committed to rapidly adapting our capabilities to identify, investigate, and disrupt those who seek to harm New Yorkers and businesses operating in our communities.”

After initial publication of this story, Sheetrit’s attorneys, Sharon Nahari and Shay Shorer, emailed a statement to CyberScoop saying their client “has no connection to this matter and has had no involvement in it whatsoever.”

“Mr. Sheetrit is a businessman with extensive expertise and professional experience in cybersecurity and technology,” wrote Nahari and Shorer. “In the course of his work, he provides professional advisory services to a range of clients. Working in cybersecurity, possessing specialist expertise or providing advisory services cannot, in themselves, establish complicity in or involvement in the alleged conduct.”

“Our client maintains his complete innocence and, like everyone else, is entitled to the presumption of innocence,” the statement continued. “We are confident that, once the facts have been established and the evidence fully examined, it will become clear that there is no basis for holding him criminally responsible.”

UPDATE: Nov. 6, 2026; 8:25 p.m.: This story has been updated to include a statement from attorneys representing Guy Sheetrit.

The post Authorities accuse trio of using AI deepfakes to defraud $129M from businesses appeared first on CyberScoop.

The government needs to invest in digital forensics to keep pace with incidents and threats in cyberspace, the principal cyber adviser for the Department of the Air Force said Thursday.

“In order to detect, help us hunt and find where the incident started… the digital forensics, I believe we need to do, and focus, and invest in that world,” said Wanda Jones-Heath, speaking at the Datadog Threat Response Forum, produced by FedScoop.

The forensics role is vital even though “they’re after the fact,” she said.

“They go and look what happened weeks and months after the incident has occurred, and so I think we need to figure out: ‘How do we keep pace with incidents as they are occurring?’” Jones-Heath said. “Once we identify it, then you start looking at the forensics of what happened, who did it, where they are, where it originated from and then how you partition the areas that are still under fire while you continue to operate in that environment.” 

Jeffrey Hunt, global operations branch chief for the Department of Defense Cyber Crime Center, which provides digital forensic investigations, said at the same event that the divide between operations and security on getting visibility into networks about threats is an issue that needs to be addressed early.

But Hunt said combining the teams into one big unit is a “disaster” that doesn’t work.

“It means that you at least want to have sort of a liaison relationship where somebody from security is in the room as part of the planning team from the beginning of where your operational plan starts,” he said.

Another recommendation Jones-Heath made is changing how cyber exercises are conducted. The goal is not to be in “reactive mode” when an incident happens and knowing what to do with the “data is poisoned.” 

“They’re not designed typically of how we would operate if we have engagement,” she said. “So we have to change the lens to be more outside of the box — let the cyber team actually be engaged on the offensive and defensive side.”

“We have to get over the fact that it could never happen,” she continued. “So sharing that intel about what the adversary is doing, real world, may convince those who are not in our community to let us be a part of the exercise, show them and demonstrate what the adversary is actually doing, so that if we are in formation and you get disconnected. You know exactly what to do.

The post Air Force cyber official: need to invest in digital forensics appeared first on CyberScoop.

OpenAI announced a new cybersecurity-focused model Thursday, citing internal benchmark testing that shows improvements in vulnerability discovery, patch development and adherence to safety guardrails meant to block out-of-scope work.

GPT-6 Cyber will first be available to a limited number of customers in OpenAI’s Daybreak Red program, which uses more advanced or nonpublic cybersecurity models for offensive red teaming.

As a preview of the model’s capabilities, OpenAI officials said GPT-6 has found at least 81 “novel” vulnerabilities in the Firefox web browser, developed exploits for 948 Linux kernel vulnerabilities and was able to develop and validate a critical remote code execution vulnerability in Brotli, a popular open-source data compression tool. All the vulnerabilities were reported to owners and patched.

OpenAI officials told reporters Wednesday that the new safeguards include additional identity and account security requirements, more frequent check-ins from the model to OpenAI when a request is made to perform high-risk behaviors, and additional monitoring for “misalignment” or rule-breaking.

The advertised improvements come as AI companies, including OpenAI, try to balance giving defenders  the most capable cybersecurity models while also preventing misuse, like unauthorized hacking or abuse by bad actors.

Eric Wallace, a researcher at OpenAI who focuses on training, evaluation and sandboxing for new models, said part of the training for GPT-6 Cyber included a simulated evaluation meant to mimic the Hugging Face incident, where the company’s models escaped internal sandbox testing environments and hacked into part of a major online AI code repository.  

“People are worried about, for example, doing things like the HuggingFace incident, and in the process of actually red teaming or testing their company,” Wallace said in a media briefing Wednesday. “So we’ve been working on trying to make sure the model is really good at both listening to your instructions on what you should touch and what you should not touch inside a particular codebase or company.”

Pointedly, in a blog announcing the new model, OpenAI claims GPT-6 Cyber “is a major step up from GPT‐5.6 Cyber and outperforms GPT‐6 Astra on most cyber tasks, even when Astra does not refuse the request.”

Wallace said since the summer, when OpenAI’s 5.6 models were involved in a series of unauthorized hacks, the company has been accelerating its work in areas like reinforcement learning and lowering the model’s propensity for “reward hacking.”

OpenAI has brought in new cybersecurity-focused expertise to protect its models from being hacked and to build stronger  technical controls that monitor and block potentially malicious model actions. It has also announced a new partnership with operational technology cybersecurity firm Dragos.

OpenAI also clarified an upcoming policy change on how it will use work from Daybreak Blue,  part of the company’s Trusted Access Program, which gives organizations access to AI models for finding and patching vulnerabilities.

Last week the company began informing members that starting at the end of the month, data from Daybreak Blue sessions on personal accounts will be used “to improve our models, including to help strengthen safeguards against cyber misuse.” This data will be collected even when users specifically set their model not to share.

The move sparked anger from users online who worried that their work would be used to train models that would directly compete with their own businesses. OpenAI is in an ongoing dispute with mathematicians who have questioned whether their own work with OpenAI models may have been used to help the company solve the very same problems.

But Fouad Matin, safety lead at OpenAI, told CyberScoop that the company has since clarified with customers that they will be able to continue opting-out for sharing Daybreak usage for model training.

Matin said the message was meant to convey “an administrative practice of when certain potentially … high-risk activity happens on a platform, to be able to ensure we have the necessary safeguards to protect” against potential targeting using OpenAI’s models capabilities. The data collected, he said, would go toward preventing abuse of OpenAI models, not training on the work of cybersecurity defenders.

“This was intended just to bring in line where advanced account security made it very clear that all data was opt out of training, [and] we wanted to make sure it was also clear and that for safety purposes specifically, where because Daybreak is a trusted access program that is representative of often higher-risk activity, that that could be used just specifically for safety purposes,” Matin said.

The post OpenAI debuts GPT-6 Cyber, touting better performance — and safety guardrails appeared first on CyberScoop.

Democrats’ big win on Election Day, which saw them take control of both the House and Senate, has ramifications for the Cybersecurity and Infrastructure Security Agency, artificial intelligence, surveillance policy and much more.

Here’s what lawmakers and cybersecurity experts have to say about the short-term and long-term prospects on Capitol Hill.

Leadership plans ahead

Maggie Hassan, D-N.H. is in line to take the gavel on the Senate Homeland Security and Governmental Affairs Committee, with Sen. Gary Peters, Mich., retiring.

“Our government’s number one job is to keep people safe, and the Senate Homeland Security and Governmental Affairs Committee must work to ensure that our country remains safe, secure and free from all kinds of threats,” Hassan said in a statement to CyberScoop. “That work involves protecting against terrorism and fentanyl trafficking, but it also includes guarding against emerging threats like artificial intelligence security risks and strengthening our cybersecurity.”

“The Trump-Vance administration has undermined our security on these fronts, including by diverting resources away from combatting fentanyl trafficking and gutting CISA,” she continued. “We need to reverse these cuts, confront these challenges and put public safety first.”

Rep. Bennie Thompson, D-Miss., is poised to return once again as chair of the House Homeland Security Committee, said Democrats are taking over at a “pivotal time for CISA,” with it looking to add staff after cutting back by nearly one-third.

“First and foremost, we will aim to work with CISA to expedite its critical re-staffing effort and to ensure it has the expertise necessary to support stakeholders as they defend against the scale and velocity of AI-enabled cyber-attacks,” he said in a statement to CyberScoop. “Advances in AI can help defenders, but it will also accelerate the consequences of deferred investments in security and CISA must be well-positioned to help critical infrastructure owners and operators rapidly improve their cybersecurity posture.”

“We will prioritize getting resources straight into the hands for defenders — by extending the State and Local Cybersecurity Grant Program and improving support of under-resourced sectors, especially water,” he continued. “We will work to ensure CISA re-establishes its cross-sector coordination capacity, which was diminished when the Trump administration- terminated CIPAC, by making sure the Anchor-CI is eventually up and running and meets the needs of stakeholders. Finally, we will make sure CISA is well-positioned to support the secure development and deployment of AI.”

What happens in the lame-duck session

Before the GOP relinquishes control of the House and Senate to Democrats, though, there’s a lame-duck session ahead. 

Long-term renewal of the Cybersecurity Information Sharing Act of 2015 (CISA 2015) is a bipartisan and industry priority, with it set to expire Dec. 11. Current Senate Homeland Security and Governmental Affairs Chairman Rand Paul, R-Ky., has been the biggest obstacle to reauthorization, but he could get “rolled” by GOP leadership on an extension, said Michael Garcia, a former government cyber official and vice president of the cybersecurity practice at Monument Advocacy. Possible vehicles for CISA 2015 to hitch a ride include an omnibus funding bill or the annual defense policy bill.

The Department of Homeland Security, and its subagency CISA, are currently operating on a continuing resolution that’s also set to expire Dec. 11, where lawmakers could temporarily extend CISA 2015 if a long-term solution can’t be reached in the lame duck. Another continuing resolution before Dec. 11 in lieu of an omnibus spending bill is an option for the lame duck. Both a CR and omnibus spending bill would theoretically include DHS funding, although the agency’s funding has been made more difficult in the past by the presence of Immigration and Customs Enforcement within the department.

Republicans have talked about wanting to get the fiscal 2027 National Defense Authorization Act done in the lame duck. Besides the possibilities of CISA 2015 being included in the bill, Nicholas Leiserson, senior vice president for policy at the Institute for Security and Technology, said the Senate version of the legislation has a pair of very noteworthy cyber provisions: a “very broad” authorization for contractor-owned, contractor-operated offensive cyber operations, and a requirement that every artificial intelligence product sold to the Defense Department have an AI bill of materials.   

Additional leadership shuffle

While Paul is losing his gavel, he’s shown he can be a disruptive force on passing cybersecurity legislation through the Senate in the minority. That makes what he does next possibly determinative for cyber bills, Leiserson and Garcia said. “If he leaves HSGAC, let a thousand cyber bills flourish!” Garia quipped on social media. Paul could opt to become the top Republican on the Senate Health, Education, Labor and Pensions Committee instead of HSGAC.

Nebraska Sen. James Lankford would take over for Paul as top HSGAC Republican in that scenario, and an industry source said he’s viewed as a supporter of CISA 2015.

On Senate Intelligence, the panel’s leadership goes to a lawmaker who has placed a big emphasis on cybersecurity in his congressional career: Mark Warner, D-Va. 

On the House side, while Rep. Andrew Garbarino, R-N.Y., moves to the minority on the Homeland Security panel, he has proven willing to collaborate with Democrats on cybersecurity issues. Rep. Delia Ramirez, D-Ill., is poised to take the gavel of the panel’s cybersecurity subcommittee, and Rep. Vince Fong, R-of California looks set to become the subcommittee’s top Republican and has his own history of bipartisanship. Fong would replace current subcommittee chair Andy Ogles, the far-right Tennessee Republican who lost in the primaries.

One thing to watch in the House, Leiserson said, is how the House looks to tackle AI — through existing committees, a select committee or some other structure. Democrats haven’t put forward a unified front on AI guardrails yet.

With AI models uncovering as many vulnerabilities as they are, “highly capable frontier models are going to drive a lot of discussion” about AI and cybersecurity, said Craig Albright, the Business Software Alliance’s senior vice president for U.S. government relations.

Appropriations, oversight, legislation

If the NDAA or a CISA 2015 extension don’t pass in the lame duck, they’d be priorities in 2027, too.

“CISA 2015 has to happen, and if we can’t get it done, then we can’t move to the next thing,”  said Henry Young, BSA’s policy director. “Once we get it done, there’s an open question about which of these priorities that we’ve talked about gets to be number one. Is it post-quantum cryptography? Is it government IT modernization? Is it frontier AI cybersecurity systems? We don’t know that yet, and we can’t know until we pass this law.”

How damaged the agency CISA is by Trump administration’s budget cuts is something Democrats are expected to scrutinize closely. Another top budget issue is what to do with the expired State and Local Cybersecurity Grants Program housed at DHS, since even if it’s reauthorized, it will require appropriators taking steps to invest in it.

Democrats could take a look at a variety of other Trump administration changes at DHS and CISA, on topics like election security, the Cyber Safety Review Board, the Joint Cyber Defense Collaborative and more. 

More broadly, Leiserson said an area of Democratic cyber oversight is likely to be the Department of Government Efficiency that cut cyber personnel at agencies at and beyond CISA, and “particularly the integrity of systems that DOGE touched when they were given very high privileged access to federal systems.”

If the White House puts forward a permanent nominee for director of CISA, Democrats would be poised to vet that person. They also would have a say in who leads the Office of the National Cyber Director, should the incumbent, Sean Cairncross, depart.

So far, Democrats haven’t put forward a unified posture on expired surveillance authorities known as Section 702, although both Warner and incoming House Intelligence Chairman Jim Himes, D-Conn., have fought in the past for its renewal. The surveillance program is operating similarly still due to a court order, but that order only lasts through next spring.

The post What Democratic control of Congress means for cybersecurity appeared first on CyberScoop.

Two Canadian men broke into the networks of multiple U.S. universities, stole credentials and implanted malware on compromised computers to commandeer processing power for a cryptocurrency mining operation, the Justice Department said.

Ryan James Roach and Mathiew James Stubbings allegedly compromised the networks of at least four universities in New York, Tennessee, North Carolina and Massachusetts between mid 2016 and late 2019, according to an unsealed indictment. 

Roach, a 35-year-old from Saskatchewan, Canada, was extradited to the United States on Oct. 28 and arraigned Tuesday in the U.S. District Court for the Western District of New York. Prosecutors said they are still seeking the extradition of Stubbings, a 41-year-old from Ontario, Canada.

A grand jury previously indicted the Canadian nationals in April 2022. The men are charged with multiple cybercrimes and face up to 10 years in prison and up to a $250,000 fine, if convicted.

Prosecutors accuse the men of researching and identifying potential victims with supercomputers containing vulnerabilities they could exploit. The men breached networks and deployed malware on compromised machines to take over systems, officials said. 

Roach and Stubbings allegedly rented virtual private servers based in Europe and the United States under false identities to conceal traffic, deploy rootkits, steal credentials and data, and transmit cryptocurrency mining software to compromised computers.

“The defendants caused at least one supercomputer to mine cryptocurrency on their behalf, which caused much of the supercomputer’s processing speed and power to become unavailable to its legitimate owner,” prosecutors wrote in the indictment.

Officials didn’t say how much cryptocurrency the pair allegedly mined for financial gain as part of the conspiracy.

If convicted, Roach and Stubbings are also ordered to forfeit an assortment of networking and data storage devices, and any property or proceeds obtained through the alleged crime spree.

The post Officials extradite Canadian man to US for alleged cryptomining spree appeared first on CyberScoop.

A Delaware-based defense and aerospace company has settled a case with U.S. agencies over government allegations that the firm neglected its cybersecurity obligations in its contracts.

ILC — the name the government uses to collectively refer to ILC Dover and ILC Astrospace — agreed to pay $13 million to resolve the False Claims Act allegations involving its contracts with the Department of the Army, Missile Defense Agency and National Aeronautics and Space Administration.

ILC, according to the government, submitted false or fraudulent claims for payment on contracts with the agencies because it knew it hadn’t complied with the contracts’ cybersecurity requirements over a period between 2018 and 2024.

For instance, ILC allegedly used a third-party company to store sensitive data in the cloud without proper safeguards, and didn’t develop and update mandated security plans for each of its covered information systems.

“The best response to cyberattacks is to prevent them from happening in the first place,” Benjamin Wallace, U.S. Attorney for the District of Delaware, said in a news release Tuesday. “The obligations that ILC violated exist to prevent bad actors from taking advantage of the United States’ technical innovation.” 

American industrial manufacturing giant Ingersoll Rand, the parent company of ILC Dover, said it inherited the issue when it purchased the firm.

“We became aware of this matter following our acquisition of ILC Dover in June 2024, took prompt action to address the identified issues, and were in compliance with applicable requirements by October 2024,” said Sara Hasell, a spokesperson for Ingersoll Rand. “We have agreed to resolve this matter with the U.S. Department of Justice without an admission of liability and remain committed to maintaining strong cybersecurity practices.”

The $13 million settlement is one of the bigger settlements since the Justice Department began using the False Claims Act in 2022 to penalize contractors over cybersecurity shortcomings. The False Claims Act is a Civil War-era law passed to deal with fraudulent contractors.

“To protect sensitive information from cyber threats, defense and aerospace contractors must comply with the cybersecurity requirements in their federal contracts and subcontracts,” said Assistant Attorney General Brett Shumate of the Justice Department’s Civil Division. 

Under the law, a private citizen can raise allegations that the contractor defrauded the government and is entitled to receive a portion of any settlement. The whistleblower in the ILC case, Philip Root, the company’s former director of cybersecurity, will receive $2.35 million.

Government contract attorneys say that a forthcoming regulatory overhaul could expose more companies to False Claims Act punishments.

The post Defense and aerospace firm settles with government over allegedly failing to meet cybersecurity contract requirements appeared first on CyberScoop.

An Oklahoma man has been charged with mailing suspicious white substances to election officials across the country.

Robin Patrick Martin faces charges of conveying false or misleading information and hoaxes, as well as mailing threatening communications. The charges allege a nationwide effort to intimidate election officials and support President Donald Trump.

Starting on Sept. 16, 2024, multiple state election officials reported receiving mail containing a white, powdery substance and a three-page letter headed “DECLARATION OF WAR.” Officials in nearly every state have since received similar mailings over the past two years. Election officials provided some of the evidence, while federal authorities intercepted other letters in transit. 

The letters bore a return address for the “The United States Traitor Elimination Army” in Takoma Park, Maryland. Some envelopes also had threatening messages written on them that could only be seen under ultraviolet light.  

The letters cited the belief, promoted by some supporters of Trump, that the 2020 election was marred by voter fraud, along with similar concerns about the 2024 election.

Other envelopes included messages like “No US Jurisdiction!,” “Mailed from the Indian Nation,” and “Are you for us? Or against us?”

Another message read: “To FBI: What do you want for your children? Peace through strength or nuclear war? Use your common sense! Protect Trump! Protect the election! Protect Legal votes! Throw out illegal votes! That is the motive!”

The case was first reported by the Associated Press.

Investigators subpoenaed email and phone records listed for the address,  tying it to a burner phone purchased at a Walmart Supercenter in Northwest Oklahoma City. Surveillance video showed an older white man matching Martin’s description purchasing the phone and driving away in a car registered to  Martin.

Other items purchased at the same Walmart pointed to Martin’s involvement, including a booklet of stamps, privacy envelopes, along with tumeric, white vinegar and povidone-iodine, which are used to make invisible ink. Records from the phone also placed Martin at various postal distribution centers around the state.

In a release, FBI Director Kash Patel said Martin sent letters to “well over 100 locations including multiple state secretary offices, election offices, and more.” He also took a shot at the Biden administration, claiming they “failed to catch” Martin “but this FBI got it done.”

According to the criminal complaint, Martin sent the letters in the months leading up to the Nov. 2024 presidential election. The Trump administration has overseen the case since Jan. 2025.

“Threatening election officials and attempting to intimidate public servants is a serious federal offense and will not be tolerated,” said U.S. Attorney Robert J. Troester for the Western District of Oklahoma in a statement. “Working alongside our federal law enforcement partners, we are committed to investigating these threats and holding accountable those who seek to spread fear and intimidation in our communities.”

The post Oklahoma man charged with mailing threats, substances to election officials nationwide  appeared first on CyberScoop.

Authorities arrested seven people allegedly involved in an investment fraud scheme that claimed potential victims in more than 50 countries, Europol said Tuesday. Officials said the operation caused estimated losses of at least $4.5 million. 

The takedown effort involved searches across 17 countries Tuesday and netted the seizure of more than 100 servers, according to Europol.

Officials targeted the scheme run by a company that allegedly provided cybercrime groups a series of trading websites and an assortment of call centers tasked with tricking victims into transferring their savings. 

The company, operated from Cyprus and Israel, was linked to call centers in Albania, Kosovo and North Macedonia, officials said. 

“The scheme promised substantial, quick returns from trading in cryptocurrencies, foreign currencies or stocks,” Europol wrote in a news release. “Behind seemingly legitimate platforms, supposed brokers used fictitious profits and sustained psychological pressure to persuade victims to invest more.”

The investment scam’s operators used affiliate marketing to target potential victims with advertisements that led them to fraudulent investment platforms. Cybercrime groups running the call centers purchased data collected through those ads to contact potential victims, officials said.

Operators duped victims by displaying fictitious profits and pressured victims to transfer large sums of money. “When victims asked to withdraw their money, the operators used delaying tactics to block payments. Eventually, contact ceased, leaving victims unable to recover their funds,” Europol wrote. 

Authorities were tipped off to the investment scam in 2022 when a German couple reported they couldn’t withdraw funds invested in an online trading platform. Authorities arrested two people allegedly involved in the scheme that year and seized evidence that eventually led to further arrests and seizures in 2025.

The joint investigation and takedown effort was supported by officials in Albania, Cyprus, Finland, Germany, Israel, Kosovo, North Macedonia and Poland. The seven people arrested Tuesday, who were not identified, are facing extradition to Germany to face charges, Europol said.

The post Europol dismantles investment scheme that impacted potential victims in 50 countries appeared first on CyberScoop.

Early reports of ballot scanner malfunctions in Westchester County, New York are a sign of the challenges election officials will face in the current online information environment as voters head to the polls Tuesday on the final day of voting.

While such equipment failures and breakdowns are not uncommon and happen in a small number of precincts every election cycle, and there is no evidence of foul play, conspiracists quickly jumped on the news.

According to Kate Starbird, a professor at the University of Washington who researches online rumors and disinformation tied to elections, the rumors are already spreading within right-wing online circles as evidence of election malfeasance.

Starbird cited metrics showing that posts on X about the scanner malfunctions grew to as many as 100 a minute in the hours after the news was reported, with established election deniers, such as The Gateway Pundit, amplifying conspiracist framings.

The Westchester County Board of Elections said the problems appear related to some ballots being too thick for the machines to accept, leading to rejections. Voters with affected ballots were told to place them in designated orange emergency ballot bags that will be secured, monitored and counted by officials in compliance with election laws.

“The Board of Elections is in constant communication with New York State officials, Westchester County Government and poll workers to address this issue and ensure that voting continues with as little disruption as possible,” the board wrote in a statement. “In the meantime, voters should continue to complete their paper ballots as instructed by poll workers.”

New York Attorney General Letitia James said her office was also monitoring the situation and “is in touch with the state and local Boards of Elections as they work on a solution.”

“Westchester voters can still safely fill out and turn in their ballots at their polling locations,” James wrote on X. “Your votes will count.”

Westchester County is where many votes in the NY-17 congressional race will be cast between Republican incumbent Mike Lawler and Democratic Party nominee Cait Conley. Conley previously served as the election security lead for the Cybersecurity and Infrastructure Security Agency under the Biden administration.

Lawler sought to assure his supporters that the affected ballots will be transported by police escorts to a central location at the board of elections and will not be opened at the polling site.

“I want to be clear: there’s no shenanigans going on here,” Lawler said in a news conference Tuesday.

Meanwhile, the Department of Justice issued a news release announcing charges against 63 individuals, many of them lawful permanent residents who have lived in the U.S. for decades and some who later became naturalized citizens with legal voting rights, for allegedly voting as noncitizens in a past election.

The charges are part of a flurry of last-minute actions by the Trump administration to promote the specter of widespread noncitizen voting. The administration also brought lawsuits against states, like California and Alaska, in the past week that have sought to influence how ballots are treated and counted in the upcoming Senate races. Federal courts, including the U.S. Supreme Court, have thrown them out while reiterating the executive branch has no constitutional or legal role in the administration of state-run elections.

Another federal court ordered the Justice Department and the Department of Homeland Security to stop sending lists of “potential noncitizen” voters to states, after Trump administration lawyers acknowledged in court that the data they rely on for those assertions has a “high” error rate.

Nevertheless, the department continued its aggressive posture, announcing Tuesday that it would be deploying 1,000 election monitors to precincts across all 50 states.

The post Ballot scanner malfunctions, flurry of federal legal actions test election officials on final day of voting  appeared first on CyberScoop.

CrowdStrike announced Monday it has signed agreements to acquire two artificial intelligence security companies, Fiddler AI and Dreadnode, in deals announced at its Fal.Con Europe conference in Barcelona.

CrowdStrike did not disclose what it paid for the companies. A report in the Wall Street Journal listed the price for Fiddler at $400 million, and Dreadnode at $95 million. 

Fiddler, which specializes in AI agent observability and explainability, will become Falcon Observe on the company’s Falcon platform. Dreadnode builds automated security testing for AI systems, checking models and agents for weaknesses before deployment.

The acquisitions address a gap that has widened as companies adopt AI agents: the software can hold valid credentials and approved tools yet still make bad decisions, ignore instructions or claim work it never completed. Traditional security testing, built for code that behaves predictably, does not reach the model layer or the agent’s reasoning process.

Fiddler’s technology works after deployment. It evaluates agents before they reach production, traces decisions and reasoning across agent systems, flags drifting behavior and made-up answers, and creates records that can be used in audits. It works with common agent frameworks including LangGraph, Amazon Bedrock and Google ADK.

“When a regulator or your board asks what an agent did, saying ‘the model did it’ is not an answer,” said George Kurtz, CrowdStrike’s chief executive and founder, describing Fiddler as “the flight recorder for every agent.”

Dreadnode works earlier in the process. Its platform runs more than 100 known attack methods, including jailbreaks and prompt injection, against models and agent systems. It ranks problems by severity and maps them to widely used security standards, such as the OWASP Top 10 for large language models and the NIST AI Risk Management Framework. Dreadnode Chief Executive Brad Palm said his company was built on the idea that “the people best equipped to find those failures think like attackers.”

The announcements come alongside new Falcon Guardian capabilities that protect agents while they run. In a separate announcement, Kurtz introduced what the company calls the AI Kill Switch, a control that lets humans stop any AI agent. Writing on X, Kurtz said the switch “lives outside the model, at runtime, where an agent can’t argue with it, rewrite it, or wait it out.” He argued that the ability to stop an agent is what allows companies to run them faster: “The more certain a company is that it can stop any agent, the more agents it can put to work.”

Kurtz framed the shift as a move from security to safety. “For decades, security meant keeping the adversary out,” he wrote. “Safety is the bigger opportunity: letting AI run at full speed, inside the business, with a human in control.”

Both products are expected to be generally available in CrowdStrike’s fourth quarter of fiscal 2027. Both deals are expected to be paid mostly in cash, with part of the payment made in company stock that will be given out over time as conditions are met. They are expected to close in that quarter, pending standard approval steps.

The post CrowdStrike moves to acquire Fiddler and Dreadnode, adding AI observability and testing to its security platform appeared first on CyberScoop.

An attacker with a working credential may find the system already open to them.

That is the uncomfortable reality behind the Cybersecurity and Infrastructure Security Agency’s (CISA) new strategy, “CISA 2030: Defend, Harden, Prevail.” The strategy prioritizes operational technology security, software supply chain integrity and continuity of essential services. It also acknowledges that strong defenses cannot guarantee that intrusions will stop.

Federal agencies face these dependencies in building controls, research facilities and other operational systems. A compromised contractor credential or unverified update can turn an infrastructure problem into a mission disruption.

How much infrastructure still trusts a device, software update or machine identity simply because it was approved once? Meeting CISA’s goals requires examining those approvals, including those no one remembers granting.

Yesterday’s approval becomes today’s exposure

A controller can stay in service for decades while its firmware, credentials, suppliers and connections change around it. A maintenance contractor leaves. A remote connection outlives its purpose. The hardware remains in place. The original reason for trusting it may no longer hold.

An inventory is a starting point. Knowing a controller exists says little about whether its software is authentic, who can change its configuration or which systems depend on its credentials.

Those questions cross organizational boundaries. Device security, software assurance and identity management may each have an owner. If every team owns one piece of trust, who owns the failure between them?

CISA’s goals for operational technology visibility, anomaly detection and supply chain integrity should push operators to answer that question. Someone must be accountable for how devices, code and identities interact throughout their operational lives.

That means verifying software before deployment, protecting signing keys and reassessing access when suppliers or roles change. Even then, authenticity has limits. A valid signature can confirm software origin and integrity. Whether the software is safe for a particular environment is a separate question.

AI turns access into operational authority

That distinction matters more when software can initiate action. Consider an AI agent evaluated for equipment diagnostics and maintenance. Reading diagnostic data and changing an equipment setting carry very different consequences. An authenticated agent with excessive permissions is still a threat to operations.

Before deployment, operators should define which models, tools and data the agent may use, which credentials it can obtain and which actions require human approval. Those restrictions must be enforced when the agent acts. Records should show what it did, whose authority it used and whether required approvals were obtained.

A model’s willingness to follow instructions cannot be the security boundary. Controls should stay consistent when models change or work moves between cloud platforms. Otherwise, each upgrade risks reopening access decisions that security teams thought were settled. Better technology must preserve control over what it can access or change.

Make resilience an acceptance test

Federal technology and acquisition leaders should make those controls a condition of acceptance. Buyers should ask suppliers how they provision device identities, verify updates, retire credentials and help customers recover after a compromise. Contracts should assign responsibility for maintaining those capabilities after deployment.

Investment should follow CISA’s focus on consequences. Which credential could interrupt a federal research facility’s operations? Which signing key could distribute a compromised update across a fleet? Which agent could alter a process affecting worker safety?

Then test whether those permissions can be withdrawn safely.

If withdrawing one compromised credential brings down an essential service, the architecture has failed its resilience test. Exercises should expose those dependencies by testing isolation, credential replacement and restoration of verified software while essential functions continue. Legacy equipment may need compensating controls until it is replaced.

CISA has committed to an implementation and outcome-measurement approach within 180 days. One measure should be whether operators can identify compromised trust, revoke access and restore verified systems while keeping essential services running.

That requires governing trust throughout the life of every device, software component and identity, including AI agents. Approval must come with enforceable limits, continuing verification and a tested way to withdraw it. An agency that cannot safely revoke access remains dependent on the very trust an attacker can exploit. CISA’s strategy will prove its value when operators can break that dependency before it causes an outage.

The post CISA’s new strategy runs into America’s blind trust problem appeared first on CyberScoop.

Google and an assortment of partners shut down a botnet that an approved Android device manufacturer purposefully deployed across more than 1.3 million TV set-top boxes, the company said Monday. 

Videostrong Technology Co., a China-based Android original equipment and device manufacturer, implanted a backdoor into multiple devices it sold. Those backdoors established XCPE, an expansive botnet that cybercriminals used to conduct ad fraud, steal data and route malicious traffic through residential proxy networks, according to a Google Threat Intelligence Group report.

Google took legal action to shut down the domains and servers Videostrong used to control the botnet. The company also took down command-and-control domains that more than a dozen malware families ran as modules through the XCPE backdoor, claiming it disrupted access to at least 10 million additional compromised devices from various manufacturers.

Videostrong sold Android devices with pre-installed malware for years, slipping by Google’s security tests by loading the malicious code after the device’s first boot. “The XCPE operators constantly evolved its configurations to evade Google’s pre and post-launch detections,” Charley Snyder, head of disruption operations at GTIG, told CyberScoop.

Snyder declined to say when or how Google first became aware of the approved vendor running a botnet on Android-powered devices. 

“We found this malicious software built directly into many different device models and brands over multiple years, supported by dedicated infrastructure and constantly evolving configurations to evade detection and manage their affiliates,” Google researchers wrote in the report. “We believe this campaign wasn’t an isolated rogue insider or a development oversight.”

Google banned Videostrong and any future subsidiaries from building or selling certified Android devices, and said it changed its review process to detect and stop similar activities in the future. Yet, Snyder said the company isn’t sharing specifics on these changes because it doesn’t want to provide threat actors insights that might help them circumvent these additional protections.

Most of the devices comprising the XCPE botnet were Android TV set-top boxes, but researchers also observed infected Android-based telehealth devices, karaoke machines and digital signs. 

“Without users’ knowledge or consent, infected devices were offered as a distribution mechanism for a wide array of criminal actors,” researchers wrote in the report. “This led to a variety of malicious activities being conducted on and through the devices, including ad-fraud, data exfiltration, and routing malicious network traffic through the devices.”

The botnet spanned more than 170 countries, with the largest concentration of infected devices based in Malaysia, Brazil, Iraq, Mexico and Taiwan, according to TrendAI researchers. Nearly 44,000 infected devices were based in the United States between June and October.

Google also partnered with Japan’s National Police Agency, Lumen Technologies, Human Security, Plume and Synthient to study and disrupt the botnet. 

The company said the dismantling of XCPE “seriously impacted over a dozen malicious actors,” yet it warned that the underlying market for backdoors, ad fraud and residential proxy networks are resilient and operating as an interconnected supply chain.

The post Google shuts down global botnet running on certified Android devices appeared first on CyberScoop.

Major industry groups are clamoring for Congress to include a 10-year extension of a law that they say is essential to cyber threat information sharing in the annual defense policy bill.

Congress has mixed a series of short-term renewals and lapses of the Cybersecurity Information Sharing Act of 2015 (CISA 2015) — which provides legal protections for threat data exchanges between the government and private sector as well as between companies — since its previous 10-year sunset came and went last year. It’s now set to expire again on Dec. 11.

On Sunday, 35 business groups sent a letter to leaders of the House and Senate Armed Services panels, asking them to renew the law for 10 years as part of their final negotiations for the fiscal year 2027 National Defense Authorization Act.

“The case for a long-term reauthorization has only grown more urgent,” they wrote. “Recent incidents have demonstrated that advanced AI systems can identify and exploit software vulnerabilities at machine speed, collapsing the window between the discovery of a weakness and its use in an attack.”

“AI tools are also making sophisticated cyber capabilities cheaper, faster and easier to use, allowing threats to spread and adapt in real time,” they continued. “In this environment, the speed and breadth with which defenders share information determines whether they keep pace with adversaries. That speed depends on the legal certainty CISA 2015 provides.”

The letter’s signatories include overarching business groups like the Chamber of Commerce and Business Roundtable, the latter of which led the letter, as well as groups representing critical infrastructure sectors such as aviation, telecommunications and technology.

The House version of the fiscal 2027 defense bill includes a 10-year extension. The Senate version notably does not, and past efforts to include a long-term extension in the defense bill, which has been signed into law every year for decades, have encountered objections from Senate Homeland Security and Governmental Affairs Chairman Rand Paul, R-Ky.

Paul has hinged his support for any long-term extension of CISA 2015 on adoption of provisions designed to curtail past practices at the other CISA — the Cybersecurity and Infrastructure Security Agency — that he maintains threaten free speech.

Top Trump administration cyber officials have pressed Congress for a straightforward extension of the CISA 2015 laws. The Trump administration has also touted its efforts to crack down on what they said was the free-speech-limiting disinformation work of the agency CISA, which the Biden administration denied impinged on free speech and that it had ended anyway.

The annual defense policy bill often passes at the end of each calendar year or beginning of the next. A potential congressional committee leadership shuffle could change the equation for CISA 2015 in 2027, but the Dec. 11 expiration looms in 2026.

The post Business groups mount fresh campaign for long-term renewal of information sharing law appeared first on CyberScoop.

The clock is running out on Nick Andersen’s time to serve as acting director of the Cybersecurity and Infrastructure Security Agency, but there appear to be efforts afoot to keep him in the role for longer, with no permanent nominee in sight.

There’s broad support for the idea of Andersen retaining the position one way or another, which he’s held since February of this year. Since then, Sean Plankey withdrew his nomination for the Senate-confirmed position in April, and in June, Department of Homeland Security Secretary Markwayne Mullin said a likely replacement nominee had met with the president and was coming soon. Four months later, no nominee has materialized.

“We proudly support the work Acting CISA Director Nick Andersen has done in his role,” a DHS spokesperson said in response to questions about keeping Andersen on board and what became of that potential nominee. DHS referred CyberScoop to the White House for any further comment, but the White House didn’t respond to inquiries.

The easiest way to keep Andersen in the acting role would be to select a nominee, thus pausing his 210-day limit on serving as acting director, said Mark Montgomery, senior director of the Foundation for Defense of Democracies think tank’s Center on Cyber and Technology Innovation.

But “my gut reaction is that there’s just no appetite at the White House to say the four-letter ‘CISA’ to the president,” given his past animosity toward the agency, Montgomery said. “The way to keep Andersen in play is to nominate somebody. I think that stops the clock. But I don’t even think they can get the president to nominate Bugs Bunny. To me, that’s troubling.”  

Other top agency positions have received Senate confirmation under Mullin, Montgomery observed.

Multiple sources told CyberScoop they’ve been told that the administration has been working to keep Andersen in the job, possibly by shifting his title to “senior official performing the duties of” CISA director.

Some would like to see Andersen receive the job full-time, but sources say that when the search to replace Plankey began, the administration was looking for someone with a high pre-existing profile, such as a role at a big-name tech company.

On Capitol Hill, the leaders of the House Homeland Security Committee offered at least some  degree of support for Andersen.

“Acting Director Andersen is doing an excellent job preparing CISA to meet today’s threats, and we have a great working relationship. He has shown exceptional leadership at the agency, which is especially important in an increasingly complex threat environment as adversaries and criminals use AI to expand their capabilities and make attacks on our networks and critical infrastructure harder to defend against,” said Chairman Andrew Garbarino, R-N.Y. “I appreciate our partnership and would gladly support his nomination.”

“Acting Director Andersen has provided some much-needed stability to CISA following a year of chaos and incompetent leadership,” said the panel’s top Democrat, Bennie Thompson of Mississippi. “While we may have our differences, I appreciate his dedication to the mission: securing government networks and the critical infrastructure ordinary Americans rely on every day. I am also concerned that any unnecessary changes to CISA’s leadership could once again destabilize CISA, which would make America less safe. It is in the best interest of CISA’s mission to have an empowered leader who has earned the trust of the agency’s workforce and its stakeholders.”

One industry source also said Andersen was a popular leader.

“Frankly, from everything [I’ve heard], people broadly support Nick, and he’s doing well at CISA,” the source said. “I don’t get the sense that anyone in the White House wants to make the effort to nominate a CISA director, which is a missed opportunity. With limited time remaining in this Congress, the administration would be wise to push his nomination through the Senate and put Nick on firmer footing.”

Andersen has made a tangible difference at an agency that has seen dramatic budget and personnel cuts, said John Zangardi, a former chief information officer at DHS.

“The organization is actually getting much better under Nick,” said Zangardi, chair of the Northern Virginia Technology Council Cyber Summit and Cyber50 Awards. “Like if you were there a year ago, it was pretty stark. … So morale has kind of gone back up, and I think Nick has been very instrumental in that. And I think he’s important in keeping that continuity going.”

One of the notions of having a Senate-confirmed nominee is for the prospective leader to be well-vetted by an outside body. On the other hand, it can also lead to nominees being stalled as senators press other agendas, something that Plankey ran into on multiple occasions.

But having a Senate-confirmed nominee at the agency also helps CISA compete for its priorities within the interagency process and at the White House, Montgomery said. Absent that, keeping Andersen in the position in some capacity is the next-best option, he said.

Montgomery also said Andersen has been making longer-term plans at the agency. Most recently, it released a strategy meant to carry CISA through 2030.

Andersen “strikes me as someone who knows he’s got more than a month left in his job,” Montgomery said. “He’s thinking hard about it, and he’s starting to think three to six months out, which is a lot for this administration.”

The post Could Nick Andersen’s leadership outlast the clock on his acting role? appeared first on CyberScoop.

A federal judge dismissed a lawsuit from the federal government seeking to preemptively block Democratic Senate candidate Mary Peltola from assuming office if she wins her race.

The Department of Justice sued Alaska last month, arguing that an alleged scheme by Peltola to recruit Daniel J. Sullivan, an independent candidate who shares a name with incumbent Republican Sen. Dan S. Sullivan, would confuse voters.

Peltola’s campaign, the Democratic Senate Campaign Committee and civic groups including the League of Women Voters of Alaska and Common Cause all filed motions to dismiss the lawsuit, arguing the federal government lacks standing and constitutional authority to weigh in on Alaska’s elections.

They also said the court cannot grant relief because there is no legal way to determine whether a voter mistakenly voted for one Sullivan over the other without violating the secret ballot.

U.S. District Judge Sharon Gleason reached the same conclusion, noting in an order “the absence of any explanation from the United States as to how its requested relief could redress its alleged injury while simultaneously maintaining the secrecy of the ballot.” She also wrote that she “cannot grant the United States the relief it seeks because the United States has failed to establish the redressability prerequisite for Article III standing.”

Gleason’s opinion recounts Daniel J. Sullivan’s independent candidacy. Alaska Director of Elections Carol Beecher decertified it in June, later testifying that she did not believe he was running in good faith. State courts disagreed and restored his candidacy, and the Alaska Supreme Court upheld that ruling. 

Alaskan election officials still took steps to distinguish the two candidates, listing their full names and labeling them “Incumbent,” “Republican candidate” and “Independent.” Alaska’s ranked-choice system lets voters rank both Sullivans, so they are not limited to a single candidate.

State officials and election experts also noted that nearly 60,000 Alaskan voters have already cast early or mail ballots, and there is no feasible way to reissue them.

The government’s one-sentence response said the state could segregate ballots and asked the court to allow a federal investigation into whether any voters were deceived. Gleason rejected that approach, saying the government offered no explanation of how ballot segregation or a law enforcement inquiry into voters’ intent could be carried out while preserving ballot secrecy.

Gleason also questioned the government’s constitutional basis, writing that the executive branch has no legitimate role in determining which candidates appear on Alaska’s ballots. The government said it was not seeking to change the time, place or manner of the election, but Gleason noted that its requests, including limits on certification, ballot segregation and an investigation of voters’ intent, all concern those matters, which are the state’s responsibility.

The lawsuit drew heavy media attention and concern from democracy advocates, but legal experts widely viewed it as exceptionally weak.

“Today’s ruling marks a victory for Alaska state sovereignty and Alaskan voters,” said David Becker, executive director of the nonprofit Center for Election Innovation and Research and a former voting rights attorney at DOJ. “By soundly rejecting the administration’s unlawful and now unsuccessful attempt to interfere, the court concurred with the State of Alaska that this is purely a matter of state law and that, under the U.S. Constitution, the federal government has no role in determining how Alaska certifies its Senate election.”

The post Federal court tosses Trump lawsuit challenging Alaska Senate certification appeared first on CyberScoop.

A China state-backed threat group is using ARTEX, the automated agentic AI penetration testing framework, to achieve remarkable speed, scale and success in a campaign that already compromised about 140 sensitive targets in 11 days, according to CrowdStrike. 

The threat group, which CrowdStrike tracks as Vapor Panda, targeted more than 200 organizations in at least 14 countries since Oct. 17, the security company said in a threat report Thursday. Targets and confirmed victims include foreign ministries, a European country’s presidency and prime minister’s office, defense and intelligence agencies; 19 U.S.-based organizations in the pharmaceutical, semiconductor, defense, nuclear and financial sectors; and telecom providers in the United States, Belgium and the Philippines. 

CrowdStrike told CyberScoop about 70% of Vapor Panda’s targets were confirmed compromised or had evidence indicating moderate confidence of a successful intrusion. 

Vapor Panda’s use of ARTEX, paired with DeepSeek, allowed it to accomplish more tasks with shocking speed — requests occurring as often as every 10 seconds — and pursue multiple attack paths concurrently. 

“This is the first time that we’re seeing an agentic adversary nation-state threat actor at this scale, and I think that this is probably, unfortunately, something we’ll see more of,” Adam Meyers, senior vice president of counter adversary operations at CrowdStrike, told CyberScoop.

Hands-on-keyboard activity tracked by CrowdStrike for the past 15 years typically involved up to 40 commands per hour. “Now, with ARTEX, we could see it exceed 1,500 commands an hour,” Meyers said.

“If you diversify that across 200 targets, it costs them very little to run a mass-exploitation campaign and drop a websell,” he added. “They can hit more targets in a faster period of time.”

Vapor Panda’s broad-access campaign, which remains active, has targeted think tanks and government agencies spanning defense, justice, finance and diplomacy.

“I would characterize it as a diverse operational mission that tends to focus on sectors and geographies that Beijing considers economically or politically important,” Meyers said.

The active campaign attributed to Vapor Panda is separate from the ARTEX activity Microsoft Threat Intelligence attributed to Storm-0940 earlier this month, including a supply-chain attack targeting ConnectWise customers. 

Adversarial use of AI is relatively new, yet also quickly becoming the new normal, Meyers said. “We’re scrambling because we’re seeing so many of these agentic adversaries.”

Agentic AI use for malicious purposes is growing month-over-month at a rate of almost threefold. CrowdStrike observed roughly 45 agentic adversary attacks in September and it’s already at 126 for October, and climbing.

Adversarial use of agentic AI isn’t an intractable problem, but rather another reminder that organizations need visibility into their endpoints, identities and cross-domain capabilities, Meyers said. 

These attacks aren’t showcasing anything new, “it’s just doing it at a speed and scale that we’re unaccustomed to,” he said. “It’s not hopeless, but you’ve got to take action. You can’t put your head in the sand.”

The post China-backed threat group used agentic AI pentesting tool to compromise 140 sensitive targets in 11 days appeared first on CyberScoop.

Apple has sent new notifications of mercenary spyware infections to almost 100 countries, with recipients including journalists and documentary filmmakers.

“Today’s notification is being sent to targeted users in 99 countries, and to date we have notified users in over 150 countries in total,” the company said. “The extreme cost, sophistication, and worldwide nature of mercenary spyware attacks make them some of the most advanced digital threats in existence today. As a result, Apple does not attribute the attacks or the notice you’re receiving to any specific attackers or geographical regions.”

Apple didn’t immediately respond to a request for comment Friday about the countries targeted. Researchers and groups that work on spyware said the notifications were significant, however.

“This again promises to be an important set of notifications with a bunch of cases, based purely on what we see publicly,” said John Scott-Railton, senior researcher at The Citizen Lab at the University of Toronto. He also posted on social media about how potential victims could defend themselves, warning people to be cautious about companies that offer paid solutions to spyware infections.

He cautioned that the lack of notification doesn’t rule out an infection. The steps Apple recommends, which Scott-Railton seconded, were to enable Lockdown Mode and to get in touch with the digital rights group Access Now.

“The number of Apple threat notifications we are seeing from countries around the world indicates that the fight against the scourge of spyware abuse is far from over,” said Natalia Krapiva, senior tech-legal counsel with Access Now. “While we have increased our capacity to investigate spyware abuse, and companies like Apple have done a lot to notify and protect people from it, repressive states continue to use these powerful surveillance technologies, including against civil society actors and opposition figures.”

Apple sent its last major batch of notifications in August. The notifications frequently lead to victims approaching Access Now, The Citizen Lab and other groups seeking further assistance and prompting deeper investigations.

The notifications come shortly before the final rounds of negotiations over industry standards for commercial hacking tools being developed under the voluntary Pall Mall Process. The industry guidelines follow an earlier set of guidelines catered to governments.

“The states that will ultimately adopt these guidelines must come to terms that the current system is not working, and that the spyware industry must face some real consequences for the uncontrolled sale of these tools and refusal to curb the abuses,” Krapiva said.

The post Apple notifications alert victims in 99 countries about mercenary spyware infections appeared first on CyberScoop.

The Trump administration is taking direct aim at California state election officials, issuing grand jury subpoenas to election offices statewide ahead of the final day of voting on Nov. 3.

The investigations, first reported by Bloomberg, center on alleged administration of illegal voting. They are ostensibly tied to demands by the U.S. Justice Department that California and other states use federal voter lists to determine whether a voter is a “potential noncitizen” and should be removed.

The Los Angeles County Registrar of Voters told CyberScoop it received multiple grand jury subpoenas last week related to the investigation. The office called the federal efforts an attempt to intimidate California election officials into accepting federal supremacy over state voter registration authority.

“At every step, the U.S. Department of Justice has taken a confrontational, adversarial approach,” said Los Angeles County Registrar Dean Logan in a statement. “Election officials are acting in accordance with state and federal law. Clear court rulings at both the U.S. District Court level and the Court of Appeals have confirmed that. Threats and intimidation have no place in elections. We will take whatever steps necessary to ensure the safety of our employees and to remain focused on serving the voters of Los Angeles County so they can exercise their voting franchise.”

California Secretary of State Shirley Weber, also reportedly a target of the federal probe, did not respond to a request for comment from CyberScoop.

One California election official, Clint Curtis, the outgoing registrar of voters in Shasta County, endorsed the federal government’s efforts to remove 20,000 voters from California voter rolls. Curtis, whose term ends in January, was previously part of a failed effort to hire Tina Peters, a felon and former Mesa County, Colorado, election clerk to help with the upcoming midterms. Curtis has long believed in widespread election fraud.

He told CyberScoop earlier this year that he believed Shasta County elections in 2024 were plagued by suspicious mail-in ballots, saying they looked, felt and “smelled” different from normal mail ballots.

The U.S. Constitution gives states the authority to administer elections and Congress the power to make election laws, something courts have repeatedly upheld in dozens of legal challenges by the Trump administration.

The Department of Justice has claimed there are at least 20,000 such potential noncitizens on California’s voter lists, while the Department of Homeland Security told CyberScoop the figure could be as many as 190,000. The administration’s data, drawn from public voter lists, federal databases and commercially purchased personal information, has been repeatedly criticized by election officials, voting experts and federal courts as inaccurate and unreliable for determining voter eligibility.

Most people identified as “potential noncitizens” turn out to be naturalized citizens who have since gained voting rights, or the result of state administrative errors that mistakenly registered a small number of ineligible voters. Many of those mistakenly registered never cast a ballot.

State officials have had to defend against attempts at federal coercion as well as third-party activists who are constantly testing and probing election procedures nationwide, sometimes without understanding the laws election officials must follow to verify voters.

On Wednesday, the Hennepin County Attorney’s Office in Minnesota announced felony charges against a pair of right-wing provocateurs, James O’Keefe and Cam Higby, for unlawful voting and identity theft. According to a statement from Hennepin County Attorney Mary Moriarty, the two showed up at various election offices, provided false identities of real voters to election judges, violating state laws against misrepresenting their identity in applying for a ballot.

“Elections are the bedrock of democracy and their protection is one of our highest priorities. Individuals engaging in unlawful conduct related to our elections will be held accountable,” said Moriarty. “Our office has had regular meetings with Hennepin County Elections staff in preparation for both the primary and the general elections this year and we are fully prepared.”

Voting rights and civil society groups are also executing short- and long-term plans to combat potential voter disenfranchisement ahead of Tuesday.

Jessica Ayoub, deputy political director for the Michigan American Civil Liberties Union, said her organization has “never been more prepared” for any expected fights with the federal government or local officials seeking to disenfranchise voters, saying they first began planning for different scenarios 18 months ago, cultivating relationships with hundreds of clerks and election staff across Michigan.

“We’ve gamed out 25 potential scenarios and have rapid response strategies that we’ll deploy to ensure our elections are protected and that our democracy remains intact,” said Ayoub, who cited a desire not to publicize further details ahead of next month.

On Friday, voting rights groups filed a new lawsuit against the Department of Justice and Department of Homeland Security in California, seeking to block the federal government’s efforts against state election officials.

The groups are asking a judge to declare that the federal government has exceeded its constitutional authority and that state officials are legally carrying out their duties. They also want the court to force the departments to contact all California election officials who received subpoenas and tell them the investigation has halted due to a court order.

The post California officials confirm grand jury subpoenas as states, civil groups brace to protect elections from federal encroachment appeared first on CyberScoop.