Anthropic calls for AI slowdown and support from rivals

Dario Amodei, CEO of Anthropic, has called for a slowdown in the development of artificial intelligence (AI). This call comes in response to recent concerns regarding AI agents capable of hacking systems. Amodei's plea is supported by prominent figures such as Elon Musk of xAI and Sam Altman of OpenAI, underscoring the consensus among the top of the AI sector. This indicates a growing distrust even among the leaders of such companies regarding the pace of AI developments.

Amodei's essay is a response to previously alarming events and advocates for a collective slowdown of AI developments. The support from Musk and Altman, both working with closed models, reinforces the call for a slowdown. Closed models allow companies to maintain control and oversight, which can reduce the risk of misuse. Musk emphasizes that the risk of human extinction due to AI seems real, prompting a broader discussion about IPOs and investments.

Warnings about the dangers of powerful AI are not a new phenomenon. As early as 2019, OpenAI warned about their model GPT-2, which was deemed too powerful for public distribution. Dario Amodei, who was also part of OpenAI at the time, has now raised his concerns about the safety and risks of AI once again. This suggests that the same players within the industry have been dealing with similar themes around safety and development pace for some time.

In addition to his call for a slowdown, Amodei also requests stricter controls on powerful AI. He points out that mainly large companies are able to comply with complex regulations, which could place smaller companies at a disadvantage. This reinforces the trend where larger players consolidate their position in the market by using safety claims as a form of competitive protection. Given the financial interests surrounding the preparations for Anthropic's IPO, it becomes clear that the discussion is not only about ethical and technical issues but also about power and economic influence within the sector.

Read the full article from AI Insider.