Normal view
-
TechRadar - All the latest technology news
- I found a hidden ChatGPT setting that changes how hard the AI thinks β and the difference surprised me
I found a hidden ChatGPT setting that changes how hard the AI thinks β and the difference surprised me
-
TechRadar - All the latest technology news
- I asked ChatGPT to turn me into a 1990s action figure β and it remembered things I'd forgotten
I asked ChatGPT to turn me into a 1990s action figure β and it remembered things I'd forgotten
-
TechRadar - All the latest technology news
- People are starting to think ChatGPT is too cheap β and that might be a problem for OpenAI
People are starting to think ChatGPT is too cheap β and that might be a problem for OpenAI
-
CyberScoop
- OpenAI: βLikelyβ Chinese influence operation tried to use ChatGPT to stir debate on data centersΒ
OpenAI: βLikelyβ Chinese influence operation tried to use ChatGPT to stir debate on data centersΒ
OpenAIβs threat intelligence team tracked what it believes are two distinct clusters of activity online from groups with ties to China and posting content seemingly designed to stoke anger around divisive topics like AI and data centers.
The first, dubbed βData Center Bandwagon,β used ChatGPT to create imagery and social media comments claiming data center buildouts were raising electricity prices for Americans.
Another used the tool to develop images and online posts characterizing tariffs as a covert means for the countries to exert control over the global technological landscape. According to OpenAI, the originating prompts directed ChatGPT to only include U.S. President Donald Trump in this content, while leaving out Chinese President Xi Jinping, who has also made use of tariffs.Β Β
In both cases, OpenAI said the operators βlikely originatedβ in China. The anti-data center content was traced to an unnamed Chinese technology company that holds multiple contracts with regional Chinese governments, and both clusters used VPNs to evade restrictions, prompted ChatGPT in simplified Chinese and asked for both English and Chinese-language outputs, all while posing as Americans on social media platforms like X and YouTube.
βThis looks like a classic example of a foreign influence operation jumping onto the bandwagon of a genuine and pre-existing domestic debate and trying to manipulate it by using fake accounts posing as Americans,β online, said Ben Nimmo, principal investigator at OpenAI and author of the report.Β
While OpenAI β which has sought to raise hundreds of millions of dollars in funding to build datacenters in the U.S. β is not a neutral party, the report does not claim that anti-data center sentiment in the country is being driven or bolstered by foreign propaganda online.
Thereβs little evidence that the campaigns got much attention outside their own amplification networks. Such engagement from third parties is an imperfect but important indicator of an influence operationβs impact. OpenAI rated the campaigns a 1 and 2 on the Bookings breakout scale, scores that indicate activity on one or more platforms but no evidence of meaningful engagement by targeted audiences.
Additionally, researchers who study state-sponsored influence campaigns say these groups are happy to latch onto and amplify genuine domestic movements or messaging so long as it serves their larger destabilization goals.
Others have suggested that piggybacking off established narratives with organic momentum β like public anger at AI and data centers β can make an influence operation appear more effective.
While AI tools can be leveraged to create such internet content at scale, they often fail to gain traction. Some images used by Chinese actors appear clunky or use overly direct messaging that display a lack of familiarity with both the English language and internet virality.
βI do want to be really clear here: this was not a case of an influence operation creating a debate,β said Nimmo. βThe debate existed already. This was an influence operation from China trying to interfere in it. We didnβt see any signs that it succeeded.β
He added that while such views are βreasonableβ and βsincerely heldβ by many participants on both sides, βwhat we donβt want to see is a covert foreign influence operation posing as Americans to try to shape it, still less a foreign influence operation using the very AI that it attacks.β
According to the OpenAI report, the actors used ChatGPT to edit work reports which contained operational security details about their social media campaigns. In them, they described their goals as βestablishing persistent and credible accounts, producing visually appealing content to expand audience reach in different regions and maintaining long term account viability by anticipating platform enforcement.β
Another report fed into ChatGPT discussed how best to leverage Facebookβs content ecosystem, groups, pages, hashtags, advertising tools, recommendation systems and reporting mechanisms, as well as strategies for evading Metaβs detection of coordinated inauthentic accounts.
The campaign around tariffs also used ChatGPT to create short comments, comics in English but also Italian, Japanese and traditional Chinese accusing the US of putting profits over loyalty to its allies. OpenAI said they were targeted by the same network on X with an influence campaign alleging a widespread user data breach that Nimmo said βnever happened.β
While OpenAI said the campaigns likely originated in China, they do not directly attribute the operations to the Chinese government or actors working on their behalf, but do note that many parts of the campaign and its tactics overlap with pre-established Chinese government propaganda campaigns online.
The post OpenAI: βLikelyβ Chinese influence operation tried to use ChatGPT to stir debate on data centersΒ appeared first on CyberScoop.
-
TechRadar - All the latest technology news
- AIs like ChatGPT fall apart in classic 'Stroop' psychological test β and that could stand in the way of achieving artificial general intelligence
AIs like ChatGPT fall apart in classic 'Stroop' psychological test β and that could stand in the way of achieving artificial general intelligence
-
TechRadar - All the latest technology news
- I made an AI clone of myself based on my Google and Reddit history β and it understood me better than I expected
I made an AI clone of myself based on my Google and Reddit history β and it understood me better than I expected
-
TechRadar - All the latest technology news
- OpenAI just quietly retired the last of the GPT-4 models β and it feels like the end of an AI era
OpenAI just quietly retired the last of the GPT-4 models β and it feels like the end of an AI era
-
TechRadar - All the latest technology news
- I asked ChatGPT to build me a realistic weekly workout for a 54-year-old body β and I actually kept doing it
I asked ChatGPT to build me a realistic weekly workout for a 54-year-old body β and I actually kept doing it
-
TechRadar - All the latest technology news
- I started asking ChatGPT one extra question β and its answers suddenly became far more useful
I started asking ChatGPT one extra question β and its answers suddenly became far more useful
-
TechRadar - All the latest technology news
- I asked ChatGPT to make my daily walks less boring and more mindful β and it changed how I see my neighborhood
I asked ChatGPT to make my daily walks less boring and more mindful β and it changed how I see my neighborhood
OpenAI heralds cybersecurity, election interference safeguard plans for 2026 midterms
OpenAI on Wednesday hailed its plans to safeguard information and aid cybersecurity defenders in the 2026 midterm elections, including work to combat deepfakes and other forms of artificial intelligence misuse.Β
The announcement builds on commitments from major tech companies in 2024, including OpenAI, to protect elections from AI-infused election interference β efforts that some thought werenβt enough. Government agencies, non-governmental institutes and others have increasingly warned about AIβs ability to have a negative impact on elections even as they advertise its potential for good.
OpenAIβs plan has five planks: spreading reliable information about voting and election results, helping with cybersecurity, watermarking deepfakes, enforcing policies that ban users from deploying its tools for election interference, and weeding out political bias in its models.
OpenAI highlighted that it has made its Codex Security agentic framework and Trusted Access for Cyber framework available to election officials, and was briefing the National Association of Secretaries of State and the National Association of State Election Directors on its tools.
βThis is an important moment for cyber defenders across industries, and we believe AI plays a critical role in hardening digital infrastructure β including systems that support elections,β the company said. βOpenAI is committed to building resilience across the infrastructure stack, including in ways that support election execution.β
Some elements of OpenAIβs plans arenβt new so much as itβs taking pieces from other announcements and putting them together in one, such as reiterating last weekβs partnership with SynthID to add watermarks to images generated with ChatGPT to assist in evaluating whether something is real or a deepfake.
One new element of Wednesdayβs announcement is that OpenAI has struck a partnership with the Associated Press on sharing election data.
One election security expert welcomed the OpenAI announcement.
βGiven the prevalence and amplification of disinformation about our elections, sometimes coming from leaders in high office, itβs always a good thing when platforms and services embrace their obligation to deliver accurate information to users,β David Becker, executive director of the Center for Election Innovation & Research, told CyberScoop. βIt appears OpenAI is doing that with this announcement. I hope other platforms embrace this responsibility as well.β
The post OpenAI heralds cybersecurity, election interference safeguard plans for 2026 midterms appeared first on CyberScoop.
-
TechRadar - All the latest technology news
- I added one sentence to my ChatGPT prompts β and suddenly the advice became way more useful for real life
I added one sentence to my ChatGPT prompts β and suddenly the advice became way more useful for real life
-
TechRadar - All the latest technology news
- I couldnβt figure out how to delete old ChatGPT images from my Library β hereβs the hidden method that finally worked
I couldnβt figure out how to delete old ChatGPT images from my Library β hereβs the hidden method that finally worked
-
TechRadar - All the latest technology news
- 5 ChatGPT hacks I wish Iβd started using sooner β they completely changed how I use AI
5 ChatGPT hacks I wish Iβd started using sooner β they completely changed how I use AI
-
TechRadar - All the latest technology news
- OpenAI is giving everyone in this country free access to ChatGPT Plus for a year
OpenAI is giving everyone in this country free access to ChatGPT Plus for a year
-
TechRadar - All the latest technology news
- ChatGPT now wants to connect up to your bank accounts β so what could possibly go wrong?
ChatGPT now wants to connect up to your bank accounts β so what could possibly go wrong?
OpenAI Rolls Out Advanced Security for ChatGPT Accounts
Advanced Account Security provides stronger login methods, more secure account recovery, shorter sessions, and training exclusion.
The post OpenAI Rolls Out Advanced Security for ChatGPT Accounts appeared first on SecurityWeek.
The state of play: Microsoft 365 and Copilot
OpenAI expands Trusted Access for Cyber program with new GPT 5.4 Cyber modelΒ
OpenAI said it is expanding its Trusted Access for Cyber program to βthousands of individuals and organizations,β who will use the companyβs technology to root out bugs and vulnerabilities in their products.
The program will also incorporateΒ GPT 5.4 Cyber, a new variant of ChatGPT that OpenAI says is specifically optimized for cybersecurity tasks. OpenAIβs goal with this release is to make advanced cybersecurity tools more widely accessible.
The company said access to the program and cybersecurity-focused model will still be governed by βstrongβ Know-Your-Customer and identity verification rules to help prevent the modelβs spread to bad actors.
βOur goal is to make these tools as widely available as possible while preventing misuse,β the company said in a blog posted Tuesday. βWe design mechanisms which avoid arbitrarily deciding who gets access for legitimate use and who doesnβt.β
OpenAIβs announcement comes one week after Anthropic rolled out Project Glasswing, a similar effort that seeks to provide major tech companies with Claude Mythos, an unreleased model that Anthropic officials have claimed is too dangerous to sell commercially.
OpenAI officials noted they publicly announced Trusted Access for Cyber program months earlier.Β They have also quietly avoided direct comparisons to Mythos, and GPT 5.4 Cyber.
Cybersecurity experts in the U.S. and UK have described Mythos as a significant improvement from previous frontier models around identifying (and potentially exploiting) cybersecurity vulnerabilities, though there remains debate and speculation about the modelβs ultimate impact on information security.Β Β
Similarly, GPT 5.4 Cyber has been finetuned for testing and vulnerability research, though OpenAI wants to make iterative improvements to the program as lessons are learned.
The company has plans to allowΒ a broader group of cyber operators to use the model to protect critical infrastructure, public services and other digital systems. The company said it is also leery of having too much influence over which industries or sectors ultimately take part in the program.
βWe donβt think itβs practical or appropriate to centrally decide who gets to defend themselves,β the blog stated. βInstead, we aim to enable as many legitimate defenders as possible, with access grounded in verification, trust signals, and accountability.β
The post OpenAI expands Trusted Access for Cyber program with new GPT 5.4 Cyber modelΒ appeared first on CyberScoop.