Elon Musk urges AI labs to test each other’s models
Elon Musk called on leading artificial intelligence laboratories, including Chinese companies, to test one another’s models before their public release. In his view, such mutual peer review could increase the likelihood of identifying safety issues in AI systems, CNBC Top News reports.
Musk’s proposal
Musk said that the mutual testing mechanism he proposed would not be a perfect solution. At the same time, he said, it would significantly increase the chances of identifying problems in models before they are released to a broad audience.
He also acknowledged that laboratories competing with SpaceXAI, which is owned by SpaceX, have not yet agreed to the proposal. Musk believes that China could probably agree to the rapid implementation of such an approach.
Amodei proposed a different oversight option in his essay: third-party evaluations involving embedded evaluators. They would be expected to verify whether AI laboratories comply with safety practices.
More current news is available on the UA.News Telegram channel Telegram.
The positions of the United States and China
The administration of US President Donald Trump opposed calls from technology executives for tighter restrictions in the field of artificial intelligence. On Monday, Trump published a series of posts on Truth Social in which he called concerns about AI a hoax and fraud.
Director of the US National Economic Council Kevin Hassett told CNBC that issues related to AI should be resolved by the private sector. According to him, the US government continues to monitor the industry and will use law enforcement measures when necessary to ensure that companies act responsibly.
Trump previously suggested that attempts to slow AI development could give China an advantage in developing the technology. Amodei called this issue the most difficult dilemma. A spokesperson for China’s Ministry of Foreign Affairs, according to Reuters’ English translation, described calls by AI company executives to slow the development of the technology as scaremongering.
