Anthropic has revised its AI safety policies amid Pentagon pressure to allow military use, emphasizing competitiveness over strict ethical commitments.
In response to growing pressure from the Pentagon, Anthropic has revised its safety policies, shifting away from its previous stance as a leader in AI safety. The company is now more focused on staying competitive in the rapidly evolving AI landscape.
Initially, Anthropic had implemented the Responsible Scaling Policy (RSP) in 2023, which included voluntary guidelines aimed at mitigating catastrophic risks associated with AI technology. This policy originally committed the company to halt AI development if it identified significant dangers.
Jared Kaplan, Anthropic's Chief Scientist, stated, “We decided that we wouldn’t help anyone by simply stopping the training of AI models. Given the rapid progress in the industry, it seemed unreasonable to take unilateral commitments while competitors move ahead.”
The updated RSP allows Anthropic to continue developing AI as long as it does not fall significantly behind its competitors. The firm emphasizes transparency in its operations, pledging to publish detailed testing results while maintaining system controls.
Meanwhile, the Pentagon is advocating for the use of AI in citizen surveillance and the creation of autonomous weaponry, a position Anthropic has opposed. High-ranking officials have indicated that without compliance with their demands by February 27, the company could face severe repercussions, including being labeled a threat to supply chains, which could jeopardize its government contracts.
CEO Dario Amodei recently met with Defense Secretary Pete Hegseth to address these issues. The Pentagon has hinted at the possibility of invoking the Defense Production Act to mandate the use of Anthropic's technology if necessary.
Franklin Turner, a government contracts attorney, remarked, “This scenario is unprecedented and will almost certainly lead to a wave of litigation if the administration takes negative actions against Anthropic.”
Despite the Pentagon's pressure, Anthropic has stated it will not relax its restrictions on military applications of its technology. Ethereum co-founder Vitalik Buterin publicly supported this stance, emphasizing the importance of maintaining ethical boundaries in AI development.
Recent reports have also highlighted the use of Anthropic's AI model, Claude, in a military operation to capture Venezuelan President Nicolás Maduro, raising further concerns about the implications of its technology in defense applications.
