Anthropic CEO Dario Amodei outlined a plan to slow AI development, emphasizing the need for caution amid rapid advancements. He said the company is 'unilaterally committing' to implementing embedded evaluators from third-party organizations like METR.
Amodei stated that AI has been advancing 'drastically faster' in recent months, particularly with its 'growing ability to build the next generation of AI.' He argued that slowing progress is essential to ensure safety and alignment.
The proposal includes giving evaluators access to company resources, such as badges, desks, and laptops, comparable to internal risk assessment teams, with exceptions for legal or contractual reasons. Amodei also called for 'common safety standards' and 'limits on the rate of unchecked AI progress' among leading AI companies.
"We must slow the pace at which we improve the capabilities of AI models," said Amodei. "Progress will still seem fast, and we must make wise use of the time we gain."
The announcement follows concerns raised by researcher Jacob Coxon, who resigned from Anthropic over fears that AI companies are 'gambling with our lives.' Amodei acknowledged the risks but emphasized the need for a 'balanced' approach to ensure AI benefits society.
Amodei did not specify how the US government should mediate safety discussions, and he raised the challenge of coordinating with authoritarian regimes. He suggested focusing on 'prohibiting certain narrow and obviously dangerous uses of AI, such as using AI for the production of biological weapons.'
Source: techcrunch
