Anthropic secretly submitted a draft of the AI Safety Act to the US Congress

In this draft, it is allowed to stop the models only if there is a possibility of death of more than 50 people or damage of one billion dollars. The post Anthropic secretly presented the draft AI safety law to the US Congress appeared first on Safir Arjan.

In this draft, it is allowed to stop the models only if there is a possibility of death of more than 50 people or damage of one billion dollars.

Become a special member of Arjan Ambassador in order to benefit from and access special facilities and different departments in Arjan Ambassador.

Anthropic representatives have apparently secretly met with lawmakers in the US Congress and presented them with a draft bill to oversee the safety and security of advanced and frontier artificial intelligence models. This issue was raised based on a report from Politico magazine and quoting 3 knowledgeable people.

According to the details announced about this draft, the government will only be allowed to stop the release or use of an artificial intelligence model if that model poses a foreseeable and imminent danger; A risk that leads to more than 1 billion dollars in financial damage or endangers the lives of more than 50 people and causes severe injuries. According to the reports, these threats should specifically result from the creation or release of biological weapons, model defiance and evasion of developer controls, or conducting cyber attacks.

An Anthropic spokesperson told Reuters that the bill includes proposals that were made in the company's "Advanced Artificial Intelligence Framework" document earlier this year. According to him, the plan gives the government "the ability to prevent and deter the dangerous release of the most powerful models" while forcing developers to test their AI frontier models, be transparent about their findings and subject the models to independent body evaluations.

According to this legal draft, independent evaluators registered with the US Department of Commerce will be required to periodically evaluate their advanced artificial intelligence labs and models. Also, artificial intelligence companies and laboratories can pay for the services of these evaluators, provided that the payment of these amounts has nothing to do with the final result of the evaluation.

If the evaluators determine that a company's models have an imminent risk than a predetermined one, they are obliged to send the report directly to the Ministry of Commerce within a maximum period of 72 hours after completing the investigations.

Even though the Ministry of Commerce can request a judge to issue an order to stop the activity of the model after examining and verifying the risks, but the text of this bill explicitly prohibits any institution, ministry, organization or other government authority of the United States from suspending the stages of development, supply or internal use of the models. Ultimately, it will be up to a federal judge in the U.S. District Court for the District of Columbia to issue an injunction to suspend and block the operation of the AI model. The US Department of Commerce has not yet responded to a request for comment.

My professional career in the world of technology goes back almost 10 years; The path that started with hardware, continued with software, and now continues as a technology reporter in Sefir Arjan. I am Javad and more than three decades have passed and in addition to the world of technology, I love movies and cinema, classic rock music and Real Madrid.