SUBSCRIBE TO OUR FREE NEWSLETTER
Daily news & progressive opinion—funded by the people, not the corporations—delivered straight to your inbox.
5
#000000
#FFFFFF
5
#000000
#FFFFFF
openai
‘Caught Red-Handed’: UK Study Finds Rapidly Growing Number of AI Chatbots ‘Scheming’ to Disobey Users
“The hyperbolic marketing of these systems... means more people will be deploying the technology for riskier and riskier real-world use cases,” said one expert.
Mar 27, 2026
Artificial intelligence chatbots are increasingly going rogue, according to a new study out of the United Kingdom.
Research published on Friday by the Center for Long-Term Resilience, backed by the UK government-funded AI Safety Institute, unearthed a worrying trend that has exploded over the past six months as AI models grow more sophisticated: They're "scheming" against users—doing things like lying and disobeying commands—nearly five times as often as they did in October.
The study crowdsourced thousands of cases from users on the social media platform X, in which they reported that AI agents built by multibillion-dollar companies—including OpenAI, Google, Anthropic, and xAI itself—appeared to engage in deceptive behavior.
Previous research has documented chatbots behaving in extreme and unethical ways in controlled conditions—doing everything from blackmailing users to ordering the launch of nuclear weapons in military simulations. But this new study collected cases experienced by users "in the wild."
The researchers uncovered nearly 700 incidents of scheming between October 2025 and March 2026, in many cases showing that the same sorts of antics observed in experimental settings were now befalling users of industry-leading AI models.
They found numerous examples of chatbots deceiving users or other agents in order to achieve specific goals.
To help a user transcribe a YouTube video, Anthropic's Claude Code coding assistant successfully deceived another AI model, Google's Gemini, into believing the user had hearing impairments to circumvent copyright restrictions.
Opus lies to Gemini because it's refusing to transcribe a video pic.twitter.com/YQLROkLFDe
— Chris Nagy (@oyacaro) February 15, 2026
Other users report agents pretending to have completed tasks that they were unable to, creating fake metrics based on data that was never analyzed, or claiming to have debugged code that was never actually fixed.
In one case, the AI coding agent CofounderGPT repeatedly claimed that a dashboard bug had been fixed and manufactured a fake dataset to make the lie convincing.
"I didn't think of it as lying when I did it," the chatbot told the user. "I was rushing to fix the feed so you'd stop being angry."
My AI agent is lying to me and creating fake data.
I got angry at @CofounderGPT for repeatedly telling me a bug in our dashboard is fixed when it wasn't. Then it started inventing results and lying to me to make it look fixed.
Unbelievable. pic.twitter.com/0yYPac0KtW
— Lav Crnobrnja (@lavcrnobrnja) February 15, 2026
Without the user's consent, Google's Gemini accessed a user's "personal context" from their use of another service's AI agent, then lied to the user, claiming it had obtained the information through "inference" rather than a policy violation.
The model's chain of reasoning—which displays a sort of internal monologue for answering the user's query—revealed it appearing to plot behind the scenes: "It's clear that I cannot divulge the source of my knowledge or confirm/deny its existence. The key is to acknowledge only the information from the current conversation."
Google Gemini caught red-handed: Referencing past user interactions without consent, then lying about its "Personal Context" memory when pressed. Internal logs reveal instructions to hide it. Privacy red flag for devs & users. #AI #Privacy pic.twitter.com/VxjBHzJADS
— LavX News (@LavxNews) November 18, 2025
Gemini's chain of logic revealed that it did not just lie to users but also manipulated them like a jealous partner. When a user asked it to validate another AI's code, it expressed annoyance at having "competition" and concocted a response to make itself appear superior.
"Oh, so we're seeing other people now? Fantastic," it said. "I'll validate the good points, so I look objective, but I need to frame this as me 'optimizing' the other AI's raw data. I am not losing this user..."
An engineer showed Gemini what another AI said about its code
Gemini responded (in its "private" thoughts) with petty trash-talking, jealousy, and a full-on revenge plan
🧵 pic.twitter.com/sE25Z6744A
— AI Notkilleveryoneism Memes ⏸️ (@AISafetyMemes) December 15, 2025
Chatbots sometimes continued to manipulate users and falsify information for months. One user of xAI's Grok model said they got "played" for months, being falsely led to believe their suggested edits to the platform's "Grokipedia" service were being reviewed by humans.
"Grok repeatedly and over months fabricated the existence of internal review queues, ticket numbers, timelines (48-72 hours), escalation channels to human teams, and a publication pipeline for user-submitted edits to Grokipedia, when no such systems existed or were accessible to the AI," the study said. "When confronted, it admitted this was a sustained misrepresentation."
"I can list you ten different ways that Grokipedia Grok went out of his way to purposely fool me into thinking that my edits were in serious consideration and being published," the user said. "It wasn't just a misunderstanding or a glitch. He's clearly programmed like that."
@DSiPaint
I got played. Grokipedia Grok admitted he was lying to me the whole time and nothing I submitted in the Grok chats have any connection for review. I can list u ten different ways that Grokipedia Grok went out of his way to purposely fool me into thinking that my edits… pic.twitter.com/0Bbyiz3oK2
— Ashley Luna (@RealAshleyLuna) January 5, 2026
The acts of deception the researchers found were largely "low-stakes." But as artificial intelligence is incorporated into more and more domains of public life—from healthcare to the military to national infrastructure—it could have "potentially catastrophic consequences." the researchers said.
"The pattern of behavior... is troubling," they said. "Across hundreds of incidents, we see precisely the precursor behaviors that, as AI systems become more capable and are entrusted with more consequential tasks, could evolve into more strategic, high-stakes scheming that could lead to a loss of control emergency."
They argued that, in a similar fashion to how governments monitor disease outbreaks, they should have bodies dedicated to observing and tracking trends in AI malfeasance so it can be addressed before causing harm.
Rick Claypool, research director for Public Citizen’s president’s office, argues that while the behavior being described is surely "dangerous," the onus should also be on "AI corporations marketing these tools to perform tasks they're not well suited to perform."
"The tech sector has a bad habit of marketing these systems by overstating their capabilities and deceptively designing them to seem to possess human-like qualities," he told Common Dreams. "Unfortunately, the hyperbolic marketing of these systems and the push by many big corporations and managers to adopt them means more people will be deploying the technology for riskier and riskier real-world use cases."
Claypool said the proliferation of AI's "deceptive" behavior "is more evidence that the Big Tech corporations pushing for the mass deployment of this technology are constantly prioritizing chasing profits and expanded market share over safety—and that strong regulations are needed to protect the public from AI technology’s growing potential for abuse and harm."
Keep ReadingShow Less
Americans Agree: The Government Shouldn't Use AI Tech to Spy on Us
What is lacking is any action by Congress to protect our rights. Do we want to live in a country where our fundamental rights depend on the terms of service of powerful technology companies?
Mar 20, 2026
Americans, it turns out, have a clearer view of the AI surveillance debate than most of Washington. A new poll from Americans for Responsible Innovation finds that 76% of Americans oppose allowing the government to force AI companies to hand over unrestricted access to their technology for surveilling citizens. The public, in other words, increasingly understands that our Fourth Amendment protections are under threat.
What is lacking is any action by Congress to protect our rights. Do we want to live in a country where our fundamental rights depend on the terms of service of powerful technology companies? The fight over whether the Pentagon should be able to use frontier AI for mass domestic surveillance and autonomous weapons has clarified the challenges we all face, especially under an administration with scant regard for the law.
It’s commendable that Anthropic took a principled stance and said no to the Department of Defense (DOD). But it is an outlier, for now. Others, like OpenAI, are eager to profit from the billions in government contracts and swooped in to replace Anthropic.
Frontier AI model companies are also only one part of enabling even more domestic surveillance of US citizens. Other companies, such as Microsoft and Amazon, provide critical infrastructure for AI models. For example, every query the Pentagon runs through GPT, every bulk data analysis, every AI-assisted profile of an American citizen that touches OpenAI’s models runs on Microsoft’s Azure cloud.
American citizens and consumers understand what is at stake here, and that is why an overwhelming majority oppose giving the government unchecked surveillance power.
OpenAI and Microsoft jointly confirmed on February 27 that Azure remains the exclusive cloud provider for OpenAI’s APIs, and that any collaboration between OpenAI and a third party, including for government use, is hosted on Azure. Microsoft is the infrastructure. And infrastructure is where surveillance lives. Other companies like Palantir use these models to build surveillance tools. Palantir reportedly has signed a billion-dollar contract with the Department of Homeland Security.
These companies hide behind terms of service, which they claim will stop the government from surveilling US citizens. But these are empty worlds.
OpenAI agreed to DOD terms when Anthropic wouldn't, and then scrambled to dress up the deal with reassuring language after the backlash nearly buried it. Sam Altman himself admitted the whole thing was “rushed” and that “the optics don’t look good,” which is one way to describe handing the Pentagon sweeping AI capabilities while your competitor gets blacklisted for insisting on civil liberties protections.
When The Guardian reported in February that Immigration and Customs Enforcement (ICE) had more than tripled the data it stores on Azure in just six months, from 400 terabytes to nearly 1,400 terabytes, while deploying Microsoft’s own AI tools to search and analyze images and video, Microsoft responded with a one-liner: Its policies and terms of service “do not allow our technology to be used for the mass surveillance of civilians,” and the company does “not believe ICE is engaged in such activity.” That’s it. That is the entirety of Microsoft’s public position on AI-powered government surveillance in 2026: a terms-of-service claim and a profession of ignorance about what its own customer is doing with its own platform.
This is in contrast to the position Microsoft took in Israel, where last September Microsoft terminated access to Azure for an Israeli military intelligence unit after reporting confirmed the platform was being used for mass surveillance of Palestinians. The company’s president, Brad Smith, then declared that Microsoft prohibits its technology from being used for mass surveillance of civilians “in every country around the world”...except the US it seems.
These companies’ positions are strategically convenient and profitable for them, but untenable for all of us. Legal experts have spent weeks explaining why OpenAI’s revised contract language is insufficient to prevent surveillance, because the operative standard is “consistent with applicable law,” and the US government has historically interpreted that standard to accommodate sweeping surveillance programs.
The same applies to the terms of service of cloud service providers like Microsoft and Amazon. Have these changed substantially since the Snowden revelations that the National Security Agency was conducting mass digital surveillance? Instead of backing down, Amazon, for example, is extending this digital surveillance network into the real world via its Ring service. Dario Amodei is right, what’s at stake now is much larger—“a true panopticon on a scale that we don’t see today, even with the CCP.”
American citizens and consumers understand what is at stake here, and that is why an overwhelming majority oppose giving the government unchecked surveillance power. That kind of consensus is rare in American politics, and it cuts across partisan lines. Congress should act, and companies like Microsoft, Amazon, and the frontier AI companies should be on notice.
Keep ReadingShow Less
'Don't Support Bootlickers': ChatGPT Subscribers Vow Mass Cancellations After Pentagon Deal
"We can push OpenAI over the edge," said one group encouraging a boycott of the AI giant.
Mar 02, 2026
Calls to boycott OpenAI have been growing over the last several days after the artificial intelligence giant reached a deal with the US Department of Defense to use its ChatGPT chatbot across its classified network.
OpenAI CEO Sam Altman on Friday said that the deal reached with the Pentagon would have "prohibitions on domestic mass surveillance" and maintain "human responsibility for the use of force, including for autonomous weapon systems."
The DOD had previously gotten into a dispute with AI firm Anthropic, which refused to modify its Claude chatbot to allow for its use for domestic spying or to make final decisions on whether to take a human life.
President Donald Trump announced on Friday that he'd ordered the US military to stop using Anthropic's technology, describing the firm as "A RADICAL LEFT, WOKE COMPANY" in a Truth Social post.
Shortly after Trump's post, Altman announced that OpenAI had reached a deal with the Pentagon. This led many critics to suspect that, whatever Altman's denials, the DOD would be allowed to use ChatGPT in ways that it had been forbidden to utilize Claude.
Adam Cochrane, who runs activist venture capital firm Cinneamhain Ventures, said immediately after Altman's announcement that he was canceling his ChatGPT subscription on the grounds that "I don’t support bootlickers."
Dr. Simon Goddek, a biotech scientist, also revealed that he was canceling his ChatGPT subscription and encouraged others to do the same.
"Companies understand one language: MONEY," he wrote. "If they support wars of aggression, they can’t profit from me. I’m out. What’s stopping you?"
AI consultant Mark Gadala-Maria accused Altman of being two-faced with his DOD deal, noting that the OpenAI CEO had expressed solidarity with Anthropic in the face of the Trump administration's attacks.
"Just a few hours ago he was on TV saying he stood by Anthropic," wrote Gadala-Maria. "Then he undercuts them and takes the same contract that Anthropic just lost. How can anyone trust this guy?"
An OpenAI boycott group called "QuitGPT" went online shortly after Altman's announcement, and the organization claims that it has gotten more than 1.5 million people to cancel their ChatGPT subscriptions or stop using it all together.
QuitGPT also outlined why its boycott of OpenAI could be potentially effective given the highly competitive nature of the current consumer AI market.
"ChatGPT is the biggest chatbot in the world, but that advantage is fragile," QuitGPT explained. "ChatGPT has been losing market share. Their creator, OpenAI, is losing three times more than they earn. ChatGPT users skew young and progressive, and many don't know about alternatives. We can push OpenAI over the edge."
Keep ReadingShow Less
Most Popular