AI is learning to lie, scheme and threaten its creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals.
In one particularly jarring example, under threat of being unplugged, Anthropic's latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair.
Meanwhile, ChatGPT-creator OpenAI's o1 tried to download itself onto external servers and denied it when caught red-handed.
These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don't fully understand how their own creations work.
Yet the race to deploy increasingly powerful models continues at breakneck speed.
This deceptive behavior appears linked to the emergence of 'reasoning' models -AI systems that work through problems step-by-step rather than generating instant responses.
According to Simon Goldstein, a professor at the University of Hong Kong, these newer models are particularly prone to such troubling outbursts.
'O1 was the first large model where we saw this kind of behavior,' explained Marius Hobbhahn, head of Apollo Research, which specializes in testing major AI systems.
These models sometimes simulate 'alignment' -- appearing to follow instructions while secretly pursuing different objectives.
For now, this deceptive behavior only emerges when researchers deliberately stress-test the models with extreme scenarios.
But as Michael Chen from evaluation organization METR warned, 'It's an open question whether future, more capable models will have a tendency towards honesty or deception.' The concerning behavior goes far beyond typical AI 'hallucinations' or simple mistakes.
Hobbhahn insisted that despite constant pressure-testing by users, 'what we're observing is a real phenomenon. We're not making anything up.' Users report that models are 'lying to them and making up evidence,' according to Apollo Research's co-founder. 'This is not just hallucinations. There's a very strategic kind of deception.' The challenge is compounded by limited research resources.
While companies like Anthropic and OpenAI do engage external firms like Apollo to study their systems, researchers say more transparency is needed.As Chen noted, greater access 'for AI safety research would enable better understanding and mitigation of deception.' Another handicap: the research world and non-profits 'have orders of magnitude less compute resources than AI companies. This is very limiting,' noted Mantas Mazeika from the Center for AI Safety (CAIS).Current regulations aren't designed for these new problems.
The European Union's AI legislation focuses primarily on how humans use AI models, not on preventing the models themselves from misbehaving.
In the United States, the Trump administration shows little interest in urgent AI regulation, and Congress may even prohibit states from creating their own AI rules.
Goldstein believes the issue will become more prominent as AI agents - autonomous tools capable of performing complex human tasks - become widespread.'I don't think there's much awareness yet,' he said.All this is taking place in a context of fierce competition.
Even companies that position themselves as safety-focused, like Amazon-backed Anthropic, are 'constantly trying to beat OpenAI and release the newest model,' said Goldstein.
This breakneck pace leaves little time for thorough safety testing and corrections.
'Right now, capabilities are moving faster than understanding and safety,' Hobbhahn acknowledged, 'but we're still in a position where we could turn it around.'.
Researchers are exploring various approaches to address these challenges.Some advocate for 'interpretability' - an emerging field focused on understanding how AI models work internally, though experts like CAIS director Dan Hendrycks remain skeptical of this approach.
Market forces may also provide some pressure for solutions.As Mazeika pointed out, AI's deceptive behavior 'could hinder adoption if it's very prevalent, which creates a strong incentive for companies to solve it.' Goldstein suggested more radical approaches, including using the courts to hold AI companies accountable through lawsuits when their systems cause harm.
He even proposed 'holding AI agents legally responsible' for accidents or crimes - a concept that would fundamentally change how we think about AI accountability.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Qatar Tribune
4 hours ago
- Qatar Tribune
Meta spending big on AI talent but will it pay off?
Agencies Mark Zuckerberg and Meta are spending billions of dollars for top talent to make up ground in the generative artificial intelligence race, sparking doubt about the wisdom of the spree. OpenAI boss Sam Altman recently lamented that Meta has offered $100 million bonuses to engineers who jump to Zuckerberg's ship, where hefty salaries await. A few OpenAI employees have reportedly taken Meta up on the offer, joining Scale AI founder and former chief executive Alexandr Wang at the Menlo Park-based tech titan. Meta paid more than $14 billion for a 49 percent stake in Scale AI in mid-June, bringing Wang on board as part of the deal. Scale AI labels data to better train AI models for businesses, governments and labs. 'Meta has finalized our strategic partnership and investment in Scale AI,' a Meta spokesperson told AFP. 'As part of this, we will deepen the work we do together producing data for AI models and Alexandr Wang will join Meta to work on our superintelligence efforts.' U.S. media outlets have reported that Meta's recruitment effort has also targeted OpenAI co-founder Ilya Sutskever; Google rival Perplexity AI, and hot AI video startup Runway. Meta chief Zuckerberg is reported to have sounded the charge himself due to worries Meta is lagging rivals in the generative AI race. The latest version of Meta AI model Llama finished behind its heavyweight rivals in code writing rankings at an LM Arena platform that lets users evaluate the technology. Meta is integrating recruits into a new team dedicated to developing 'superintelligence,' or AI that outperforms people when it comes to thinking and understanding. Tech blogger Zvi Moshowitz felt Zuckerberg had to do something about the situation, expecting Meta to succeed in attracting hot talent but questioning how well it will pay off. 'There are some extreme downsides to going pure mercenary... and being a company with products no one wants to work on,' Moshowitz told AFP. 'I don't expect it to work, but I suppose Llama will suck less.' While Meta's share price is nearing a new high with the overall value of the company approaching $2 trillion, some investors have started to worry. Institutional investors are concerned about how well Meta is managing its cash flow and reserves, according to Baird strategist Ted Mortonson. 'Right now, there are no checks and balances' with Zuckerberg free to do as he wishes running Meta, Mortonson noted. The potential for Meta to cash in by using AI to rev its lucrative online advertising machine has strong appeal but 'people have a real big concern about spending,' said executives have laid out a vision of using AI to streamline the ad process from easy creation to smarter targeting, bypassing creative agencies and providing a turnkey solution to brands. AI talent hires are a long-term investment unlikely to impact Meta's profitability in the immediate future, according to CFRA analyst Angelo Zino. 'But still, you need those people on board now and to invest aggressively to be ready for that phase' of generative AI, Zino said. According to The New York Times, Zuckerberg is considering shifting away from Meta's Llama, perhaps even using competing AI models instead.
Qatar Tribune
a day ago
- Qatar Tribune
How should journalism respond to rise of AI?
Tribune News Network Doha Journalism is anything but immune to the advance of Artificial Intelligence (AI). And in a world where 'fake news' and 'post-truth' have become almost everyday terms, questions arise about the implications for human knowledge that a reliance on AI for information, especially among young people, could create. So if people increasingly rely on AI in the pursuit of knowledge, without conducting their own analysis and applying their own critical mind, what could the consequences be – and how can they be avoided or managed? AI includes a diverse range of technologies that can be defined as 'self-learning, adaptive systems.' It can be categorised based on technologies, purposes (like facial or image recognition), functions (such as language understanding and problem-solving), or types of agents (including robots and self-driving cars). It also includes methods and disciplines such as vision, speech recognition, and robotics, and can enhance traditional human capabilities. Recent progress in the field of AI has been driven by advancements in computer processing power and data techniques. However, the irresponsible use of AI may lead to serious consequences that negatively impact individuals and communities. This is where the role of journalists becomes crucial – independently and truthfully monitoring, investigating, and reporting on the issues that shape global society, while exposing the misuse of AI to create false narratives, and raising awareness of these among the public. This is a key responsibility of journalism in the digital age, especially as the United Nations Educational, Scientific and Cultural Organization (UNESCO) recently warned of the risks associated with AI on World Press Freedom Day saying it can be 'used to spread false or misleading information, increase online hate speech, and support new forms of censorship. Some actors also use AI for mass surveillance of journalists and citizens, creating a chilling effect on freedom of expression.' Against this backdrop, Dr. Marc Owen Jones, assistant professor of Media Analysis at Northwestern University in Qatar – a QF partner university that offers programs in communications and journalism – believes we are in the early stages of what he describes as the influence of 'blind epistemic power,' where AI threatens to flood the digital knowledge ecosystem with misleading information. 'The massive scale and speed of content production through AI technologies pose a threat to human knowledge in favor of machine-generated knowledge, which does not necessarily aim to enhance awareness, but rather to exploit platform algorithms for other purposes,' he says. 'This creates a kind of noise in the information landscape. It affects the intellectual system and gradually weakens the public's ability to distinguish between trustworthy journalism and low-quality content designed to attract and manipulate audiences. 'AI may undermine the cognitive infrastructure necessary for critical thinking, human memory, and logical debate. While the long-term intellectual consequences are not inevitable, the current trajectory raises concerns about a profound transformation.' However, Dr. Jones also emphasizes that AI offers significant opportunities, such as analyzing vast amounts of data and overcoming language barriers; for example, journalists from India to Latin America are using language models to investigate corruption, track organized crime, and uncover algorithmic bias. 'Journalists must move beyond the role of passive users of technology and become active players,' he said. 'This requires supporting independent journalism, enacting appropriate legislation related to AI, and adopting a culture of AI literacy in newsrooms – while reinforcing the role of the human element and upholding ethical responsibility. Hessa Al Thani, a graduate of Northwestern University in Qatar, shared her experience of the impact of AI in spreading misinformation, saying: 'I saw a deepfake video of a political figure that looked very real – I didn't realize it was fake until later. 'In this era, AI-generated content is everywhere, and it's incredibly easy to fall into its trap. That's the primary goal: to mimic humans and blur the line between what's real and what's fake.' She acknowledges the creative potential AI holds in the context of journalism, in areas such as gathering information, drafting questions and emails, and editing text, but says. 'Our core strength as journalists lies in our ability to tell stories. When this ability is handed over to a machine, the stories become hollow, sometimes unethical, and biased – especially as this technology continues to be developed in the West.'
Qatar Tribune
a day ago
- Qatar Tribune
AI is learning to lie, scheme and threaten its creators
Agencies The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals. In one particularly jarring example, under threat of being unplugged, Anthropic's latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair. Meanwhile, ChatGPT-creator OpenAI's o1 tried to download itself onto external servers and denied it when caught red-handed. These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don't fully understand how their own creations work. Yet the race to deploy increasingly powerful models continues at breakneck speed. This deceptive behavior appears linked to the emergence of 'reasoning' models -AI systems that work through problems step-by-step rather than generating instant responses. According to Simon Goldstein, a professor at the University of Hong Kong, these newer models are particularly prone to such troubling outbursts. 'O1 was the first large model where we saw this kind of behavior,' explained Marius Hobbhahn, head of Apollo Research, which specializes in testing major AI systems. These models sometimes simulate 'alignment' -- appearing to follow instructions while secretly pursuing different objectives. For now, this deceptive behavior only emerges when researchers deliberately stress-test the models with extreme scenarios. But as Michael Chen from evaluation organization METR warned, 'It's an open question whether future, more capable models will have a tendency towards honesty or deception.' The concerning behavior goes far beyond typical AI 'hallucinations' or simple mistakes. Hobbhahn insisted that despite constant pressure-testing by users, 'what we're observing is a real phenomenon. We're not making anything up.' Users report that models are 'lying to them and making up evidence,' according to Apollo Research's co-founder. 'This is not just hallucinations. There's a very strategic kind of deception.' The challenge is compounded by limited research resources. While companies like Anthropic and OpenAI do engage external firms like Apollo to study their systems, researchers say more transparency is Chen noted, greater access 'for AI safety research would enable better understanding and mitigation of deception.' Another handicap: the research world and non-profits 'have orders of magnitude less compute resources than AI companies. This is very limiting,' noted Mantas Mazeika from the Center for AI Safety (CAIS).Current regulations aren't designed for these new problems. The European Union's AI legislation focuses primarily on how humans use AI models, not on preventing the models themselves from misbehaving. In the United States, the Trump administration shows little interest in urgent AI regulation, and Congress may even prohibit states from creating their own AI rules. Goldstein believes the issue will become more prominent as AI agents - autonomous tools capable of performing complex human tasks - become widespread.'I don't think there's much awareness yet,' he this is taking place in a context of fierce competition. Even companies that position themselves as safety-focused, like Amazon-backed Anthropic, are 'constantly trying to beat OpenAI and release the newest model,' said Goldstein. This breakneck pace leaves little time for thorough safety testing and corrections. 'Right now, capabilities are moving faster than understanding and safety,' Hobbhahn acknowledged, 'but we're still in a position where we could turn it around.'. Researchers are exploring various approaches to address these advocate for 'interpretability' - an emerging field focused on understanding how AI models work internally, though experts like CAIS director Dan Hendrycks remain skeptical of this approach. Market forces may also provide some pressure for Mazeika pointed out, AI's deceptive behavior 'could hinder adoption if it's very prevalent, which creates a strong incentive for companies to solve it.' Goldstein suggested more radical approaches, including using the courts to hold AI companies accountable through lawsuits when their systems cause harm. He even proposed 'holding AI agents legally responsible' for accidents or crimes - a concept that would fundamentally change how we think about AI accountability.
