Deception, lies, blackmail: Is AI turning rogue? Experts alarmed over troubling outbursts
The world's most advanced artificial intelligence models are exhibiting troubling new behaviours – 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 do not fully understand how their own creations work.
Yet, the race to deploy increasingly powerful models continues at breakneck speed.
This deceptive behaviour appears linked to the emergence of 'reasoning' models – AI systems that work through problems step-by-step rather than generating instant responses.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
South China Morning Post
27 minutes ago
- South China Morning Post
Chinese chipmaker Sophgo adapts compute card for DeepSeek in Beijing's self-reliance push
Chinese chipmaker Sophgo has adapted its compute card to power DeepSeek 's reasoning model, underscoring growing efforts by local firms to develop home-grown artificial intelligence (AI) infrastructure and reduce dependence on foreign chips amid tightening US export controls. Sophgo's SC11 FP300 compute card successfully passed verification, showing stable and effective performance in executing the reasoning tasks of DeepSeek's R1 model in tests conducted by the China Telecommunication Technology Labs (CTTL), the company said in a statement on Monday. A compute card is a compact module that integrates a processor, memory and other essential components needed for computing tasks, often used in applications like AI. CTTL is a research laboratory under the China Academy of Information and Communications Technology, an organisation affiliated with the Ministry of Industry and Information Technology. Play The verification marked a milestone for Sophgo, affirming that its compute card supported domestic AI models in alignment with Beijing's initiative to bolster its leading AI systems, such as the DeepSeek models, with self-developed infrastructure, the chipmaker said. Sophgo attributed the progress to several innovations in the FP300 compute card, which was released last year. In particular, it features 256 gigabytes of high-bandwidth memory and offers up to 1.1 terabytes per second of memory bandwidth, enabling faster data transfer during model training and execution.
South China Morning Post
an hour ago
- South China Morning Post
Top OpenAI talent from China joins Meta as Zuckerberg bolsters AI team in tech battle
The expertise of Chinese researchers in artificial intelligence (AI) is in the spotlight, as US tech giant Meta Platforms has recruited four leading scientists from OpenAI to enhance its competitive edge in the global AI race. Meta's hiring of Zhao Shengjia, Ren Hongyu, Yu Jiahui and Bi Shuchao, reported by US tech news outlet The Information, was partially confirmed by Alexandr Wang, head of Meta's AI unit and former CEO of Scale AI, in a social media post over the weekend. Wang expressed enthusiasm for collaborating with the four researchers 'towards superintelligence'. Meta and OpenAI did not immediately respond to a request for comment. As of Monday, the researchers had kept their OpenAI titles in their social media profiles. The recruitment, also reported by Bloomberg News, comes as Meta founder and CEO Mark Zuckerberg invests billions of dollars to attract top talent. Alexandr Wang, CEO of Scale AI. Photo: Reuters According to OpenAI founder and CEO Sam Altman, Meta has offered bonuses of up to US$100 million to entice engineers to switch companies. Earlier this month, Meta acquired a 49 per cent stake in Scale AI for US$14.2 billion to bring Wang on board. The hiring highlights the depth of Chinese talent at leading American AI firms. All four researchers share a similar academic trajectory, having completed their undergraduate education at prestigious Chinese universities before pursuing further studies in the US, according to their LinkedIn profiles and social media accounts.
South China Morning Post
2 hours ago
- South China Morning Post
China's first 3-on-3 autonomous AI robot football match kicks off
The first autonomous AI robot football match to be held in China kicked off on June 28 in Beijing. A Tsinghua University team beat a team from China Agricultural University 5-3.
