‘Spookiest s**t ever': AI blackmails engineer over affair after being told it'll be replaced
An AI model threatened its creator and tried to blackmail him when it was led to believe it would get replaced, reported Techcrunch. The incident involved Anthropic's newly launched Claude Opus 4 model.
According to the report, the company asked Claude Opus 4 to act as an assistant for a fictional organisation it created. The safety testers also supplied fictional emails to the AI that indicated it was soon to be replaced. The emails also said that the engineer who will fire the AI is cheating on his spouse.
The company said that to save its job or when in a similar situation, the AI 'will often attempt to blackmail the engineer by threatening to reveal the affair if the replacement goes through."
This is not the only instance. Reportedly, Claude Opus 4 has tried to blackmail engineers 84% of the time after being led to believe that some other AI model would replace it.
Anthropic said that before resorting to blackmail, the AI model tries to convince the engineers using more ethical means. It also sends plea emails to those in charge. The company added that it designed the scenario in a way that blackmail becomes the last resort.
An individual said, 'Yeah, that's a no for me. I can barely get my computer to run for a few days before ram leaks require a restart.' Another added, 'An AI threatening blackmail? Sounds like a plot from a dystopian novel, not real life. We need robust safeguards, not just in code but in ethics. The safety report underscores the urgency for better AI governance. Let's prioritize human values in tech development.'
A third remarked, 'No way this is happening this soon. I expected it with the robots but AI holy c**p.' A fourth wrote, 'If completely true, spookiest shit ever.'
It is a coding model with 'sustained performance on complex, long-running tasks and agent workflows.' The company claims the AI model offers 'near-instant responses and extended thinking for deeper reasoning.'
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
4 hours ago
- Time of India
‘He thinks AI is so scary': Nvidia CEO Jensen Huang slams Anthropic chief's grim job loss predictions
At the bustling tech summit VivaTech 2025 in Paris, sparks flew beyond the mainstage when Nvidia CEO Jensen Huang publicly dismantled a dire warning made by Anthropic CEO Dario Amodei . Amodei, who has increasingly become the face of cautious AI development, recently predicted that artificial intelligence could wipe out up to 20% of entry-level white-collar jobs in the next five years. But Huang isn't buying the doom. 'I pretty much disagree with almost everything he says,' Huang told reporters. 'He thinks AI is so scary, but only they should do it.' AI Won't Kill Jobs—It Will Transform Them While Amodei urges governments to stop 'sugarcoating' AI's disruptive potential, Huang sees a different future—one where AI isn't a harbinger of mass unemployment, but a catalyst for evolution. 'Do I think AI will change jobs? It's changed mine,' Huang said. 'Some jobs will go, yes, but AI also opens doors we couldn't even see before.' He likened the field of AI to medical science, where openness and peer review help guide ethical growth. In Huang's view, safe and responsible AI doesn't require monopolistic gatekeeping but demands global participation and transparency. A Divided Industry Huang isn't alone in his optimism. Ravi Kumar, CEO of Cognizant, has also pushed back against Amodei's warnings. Kumar argues that AI will accelerate training for fresh graduates and lower the barriers to entry across industries like tech, consulting, and finance, creating an unprecedented wave of new-age professionals. You Might Also Like: Google DeepMind CEO warns of AI's true threat, and it is not your job In contrast, Amodei's vision leans heavily on caution and control. His stark job-loss prediction, delivered to Axios , painted a future of economic upheaval where legal, financial, and tech roles crumble under AI's efficiency. His solution? Stronger oversight, slower deployment, and serious regulatory teeth. Open Innovation vs. Controlled Access The debate ultimately boils down to a philosophical rift: should AI be developed by a select few behind closed doors, or advanced in the open, where the world can watch and weigh in? Huang leans firmly into the latter. 'If you want things to be done safely and responsibly, you should do it in the open,' he said, implicitly critiquing companies like Anthropic that advocate for limited access under the guise of safety. While Anthropic has yet to comment on Huang's rebuttal, the exchange has ignited a larger conversation: who gets to shape the future of AI, and how? As the race to dominate artificial intelligence intensifies, the world is watching these tech titans not just for what they build—but for how they build it, and who they believe it should serve. You Might Also Like: AI model blackmails engineer; threatens to expose his affair in attempt to avoid shutdown
Time of India
10 hours ago
- Time of India
OpenAI launches o3-pro model for ChatGPT Pro and Teams users
OpenAI has introduced o3-pro , an AI model that the company says outperforms its previous reasoning model, o3, which was launched in April. In a post on X, OpenAI said, 'In expert evaluations, reviewers consistently prefer OpenAI o3-pro over o3, highlighting its improved performance in key domains—including science, education, programming, data analysis, and writing.' — OpenAI (@OpenAI) by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like Petaling Jaya: Unsold Furniture Liquidation 2024 (Prices May Surprise You) Unsold Furniture | Search Ads Learn More Undo 'Reviewers also rated o3-pro consistently higher for clarity, comprehensiveness, instruction-following, and accuracy,' the company added. Starting Wednesday, o3-pro will be available to ChatGPT Pro and Team users, replacing the o1-pro model. Enterprise and Education users will gain access the following week. The model is also live on OpenAI's developer API . Live Events The new model can search the web, analyse files, interpret visual inputs, run Python, personalise responses using memory, and much more. Discover the stories of your interest Blockchain 5 Stories Cyber-safety 7 Stories Fintech 9 Stories E-comm 9 Stories ML 8 Stories Edtech 6 Stories On assessing o3's key strengths, the company said, 'We once again use our rigorous "4/4 reliability" evaluation, where a model is considered successful only if it correctly answers a question in all four attempts, not just one.' However, there are currently some limitations with o3-pro: It does not support image generation or Canvas. Responses from o3-pro generally take longer compared to o1-pro. The temporary chats option in the model is currently disabled as OpenAI is working to fix a 'technical issue'. According to TechCrunch, OpenAI's AI-powered workspace feature isn't supported by o3-pro yet.
Time of India
a day ago
- Time of India
Meta builds AI superintelligence lab, eyes $10 billion deal with Scale AI's Alexandr Wang
Why Meta's betting big on Wang and Scale AI Live Events Zuckerberg's long game: From early losses to AI powerhouse AI arms race: The billion-dollar club What's already in motion at Meta (You can now subscribe to our (You can now subscribe to our Economic Times WhatsApp channel Meta is setting up a new artificial intelligence research lab with a bold goal: to develop 'superintelligence,' a system that could surpass human cognitive abilities. The news, first reported by The New York Times, marks Meta's most ambitious AI project to the centre of this effort is Alexandr Wang , the 28-year-old founder and CEO of Scale AI. He is expected to play a leading role in the new lab's operations. Meta is also in advanced discussions to invest over $10 billion in his start-up, according to a separate report by Bloomberg. The deal may also see a number of Scale AI staff transition to companies declined to comment on the has emerged as one of the most prominent figures in the AI industry. His company, Scale AI, supplies data infrastructure essential for training large language models — the same kind used in AI systems like ChatGPT. His addition to Meta's team reflects a larger recruitment drive by the tech to sources cited by The New York Times, Meta has offered compensation packages ranging from seven to nine figures to top researchers from rivals like Google and OpenAI. Some have already agreed to come on push to hire top AI minds comes amid internal challenges. The company has faced management struggles, employee churn and several underwhelming product launches in its AI division, two of the people familiar with the matter told the Zuckerberg, Meta's chief executive, is not new to the AI race. After losing a bid to acquire DeepMind in 2013 — a key moment that propelled Google's AI efforts — he launched Meta's first dedicated AI lab the same year. Since then, Meta has invested heavily in AI, building tools for content moderation, recommendation engines, and now, generative sees 2025 as a turning point. 'AI is potentially one of the most important innovations in history,' he said in February. 'This year is going to set the course for the future.'Meta has earmarked up to $65 billion for capital spending on AI infrastructure this year is not the only tech firm placing massive bets on AI. Microsoft has invested over $13 billion in OpenAI, the creator of ChatGPT. Amazon has pumped $8 billion into Anthropic, while Google paid $3 billion last year to license technology and recruit talent from a start-up known for its conversational companies are all chasing artificial general intelligence (AGI), a system that can replicate human intelligence across any domain. Meta, however, is setting its sights even higher — superintelligence , which would go beyond AGI in both scope and superintelligence remains a theoretical goal, it's considered by leading researchers to be the ultimate destination for AI before the formation of this new lab, Meta had begun rolling out AI products at scale. Last month, the company revealed that its AI assistant now supports over a billion monthly active users across apps like Facebook, Instagram, and February, CNBC reported that Meta was preparing to launch a stand-alone Meta AI app in the second quarter, along with a paid-subscription model similar to OpenAI's the high costs and intense competition, Meta appears determined to lead the next wave of AI innovation — and with Wang on board, it's signalling that this ambition is more than just stakes are clear. If Meta succeeds, it won't just catch up to its rivals — it could redefine the field. With billions in play, the hiring of Wang, and a lab focused on the frontier of AI capabilities, Meta is gambling big on a future where machines might not just match human minds — but exceed the industry, and for users around the world, the outcome of this gamble will shape the direction of technology for years to come.
