Anthropic CEO called China the main obstacle to slowing AI development
- Anthropic CEO considers China the main obstacle to AI regulation
- He proposed a global 'speed limit', but doubts its implementation
- Chinese startups are accelerating AI development, competing with OpenAI and Anthropic
- OpenAI and Anthropic are preparing for an IPO, but the dates have not been announced yet
During the interview on the program Sunday Morning On CBS, Dario Amodei, CEO of Anthropic, noted that efforts to establish international restrictions in the field of artificial intelligence face a serious geopolitical challenge. In his opinion, the main problem is the lack of willingness of competing countries, primarily China, to comply with such restrictions in order to maintain strategic advantage.
The idea of a 'speed limit' and its realism
Amodai proposed creating a global consensus on a 'speed limit' for the development of AI technologies, allowing for a reduction in risks associated with the potentially catastrophic use of the most powerful models. However, he expressed skepticism about the feasibility of this idea, pointing to the strong incentives for technological advancement that drive states and corporations.
Despite doubts, Amodei emphasized that attempts to negotiate are necessary. He reminded of his previous calls for companies to slow down the improvement of advanced models, which was a response to growing concerns within the industry about potential threats.
Internal disagreements at Anthropic have intensified following the departure of employee Jacob Coxon, who explained his decision as a fear that companies, including their main competitor OpenAI, are 'playing roulette with our lives.'
Meanwhile, Chinese startups are actively developing their neural networks, creating competition for market leaders – OpenAI and Anthropic. At the BRICS summit in India, Chinese President Xi Jinping called for the formation of a global AI governance mechanism based on consensus, and Chinese authorities announced plans to support the open AI developer community within a bloc uniting 11 countries, including Russia.
Anthropic and OpenAI are preparing for potential initial public offerings, but there are no official dates yet. Sam Altman, head of OpenAI, stated that the company will likely postpone the IPO until next year, partly due to security concerns. Despite previous disagreements, Altman supported Amodei's initiative, stating the need to 'slow down the pace of advanced systems development.'
The initiative was also supported by well-known industry figures: Demis Hassabis, chairman of the board of Google DeepMind, noted that Amodei's essay points in the right direction, although the details require further elaboration; Elon Musk, head of Tesla and SpaceX, expressed agreement with the proposal to slow down the development of advanced AI systems.
The situation highlights the growing tension between the pursuit of technological progress and the need to ensure safety and control over powerful AI tools. While a global agreement remains elusive, dialogue between countries and companies continues to be a key element in preventing potential threats.
Source: 3DNews
