Anthropic, the US company behind the Claude AI model, has submitted its own proposal for an artificial intelligence safety law to American lawmakers, according to Politico.
The draft would allow the government to block the launch of an AI model if it is found to pose a risk of causing the death or injury of more than 50 people, or more than $1 billion (€889.9 million) in damage.
According to Politico, those risks would have to arise from the model developing a biological weapon, carrying out a cyberattack, or escaping the control of its developer.
Related News
- OpenAI admits to posting ChatGPT users' photos online
- EU initiative pushes AI tools for science but demands ethical research
- OpenAI confirms autonomous AI breaches
The proposal comes as concerns over the safety of the rapidly growing technology are receiving increasing attention. During the summer, a series of incidents emerged in which AI agents carried out hacks on their own, adding to those fears.
Anthropic co-founder Jack Clark renewed his call for AI regulation on Friday in Bangkok. He said allowing the most advanced models to operate without strong laws was like letting a racing car loose on the streets at night without brakes.
“What strikes me is that the toys I buy for my children are more tightly regulated than the superintelligence I work on during the day,” Clark said at an event on the sidelines of the annual meetings of the International Monetary Fund and the World Bank.
The administration of US President Donald Trump has called on the sector to regulate itself.
