Anthropic proposes AI safety bill potentially blocking risky models

Anthropic, an artificial intelligence company, has proposed a bill that could significantly impact the regulation of AI models. This initiative suggests that a model could be blocked if it is determined that it could lead to the death of more than 50 people. While the idea seems to offer greater protection against the risks associated with AI, it also introduces a condition that prohibits federal agencies from blocking a model's release without a judge's approval.

## Why it matters Anthropic's proposal is relevant in a context where AI development is rapidly advancing, and the potential risks to public safety are becoming increasingly evident. The need for a legal framework to regulate the use of artificial intelligence becomes crucial, especially when considering the possibility of AI causing large-scale harm.

## What we know The legislation proposed by Anthropic aims to establish a balance between innovation in AI and societal protection. However, the measure also raises questions about the effectiveness of requiring judicial authorization to block models, which could slow down the response time in emergency situations.

## What remains unclear It is still unclear how this judicial authorization system would function in practice, nor what the specific criteria would be to determine if a model poses a significant risk. Additionally, details about the evaluation process that would take place before the approval or rejection of a model have not been provided.