AI stocks tumble after US tech leaders warn of development risks

Shares in companies linked to artificial intelligence fell sharply on Monday after executives developing the most advanced AI models warned that the pace of development should be slowed to avoid potential threats to humanity.

Futures for the Nasdaq mini index dropped 1.3% during Asian trading. Shares of SoftBank, an investor in OpenAI—the developer of ChatGPT—fell as much as 13.2% in Japan.

Dario Amodei, CEO of Anthropic, called in a lengthy post on X on Saturday for AI companies to slow the development of model capabilities, citing growing concerns about AI misuse.

Elon Musk, who leads xAI, and Sam Altman, CEO of OpenAI, both said they agreed with Amodei.

Amodei wrote that within six to twelve months, AI agents could potentially control the entire internet, causing damages amounting to hundreds of billions of dollars.

Altman, CEO of OpenAI, also said the company would not proceed with an initial public offering this year, citing safety concerns.

In Japan, memory chipmaker Kioxia fell as much as 9.8% early in the session, while Tokyo Electron, a key player in the chip supply chain, dropped 3.7%.

In Taipei, shares of Taiwan Semiconductor Manufacturing Company fell 1.2%, while SK Hynix in South Korea dropped 5.3% and Samsung Electronics declined 3.7%.

Takayuki Miyajima, chief economist at Sony Financial Group, wrote in a note: "AI and semiconductor stocks in Tokyo are likely to face selling pressure after a series of weekend statements calling for a slowdown in AI development. In addition, ongoing uncertainty in the Middle East continues to weigh on investor sentiment."

In Shanghai, memory chipmaker CXMT fell as much as 3.6%, while Semiconductor Manufacturing International Corporation dropped 2.6%.

In Hong Kong, shares of Zhongji Innolight fell as much as 6.7%, while Minimax dropped up to 7.8%. Shares of ZAI, developer of the GLM AI model series, fell as much as 10.5% after a discounted share offering.

San Francisco-based Anthropic released a threat intelligence report on Thursday detailing how various actors have used its Claude AI models for activities ranging from weapons development and cyber operations to surveillance and fraud.

Concerns over the potential harms of AI intensified after Jacob Coxon, a researcher at Anthropic, resigned, saying "the people building AI seriously believe it could kill us all by the end of the decade."

Altman, OpenAI's CEO, said in an interview that "the risks AI poses to human survival are unacceptable."

While several US lawmakers have raised concerns about the rapid progress of AI and called for new regulations, former US President Donald Trump on Sunday downplayed these fears, describing AI critics as "very negative forces" promoting unlikely scenarios, and said he wants the US to remain a leader in the industry.

AI-related investments have driven much of the global stock market gains since OpenAI launched ChatGPT in 2022, but recent cyberattacks by rogue AI agents and public opposition to data center construction have fueled resistance to the industry's growth.