Anthropic and OpenAI Adjust Safety Policy Statements Amid Accelerating AI Competition

Odaily Planet Daily reports that Anthropic has removed a core safety commitment from its responsible expansion policy, no longer promising to pause training if risk mitigation measures are not fully in place. Anthropic Chief Scientist Jared Kaplan told TIME that in the context of rapid AI development, it is meaningless to make a unilateral commitment to halt training if competitors continue to push forward.

OpenAI has also revised its mission statement, removing the word “safely” from its 2024 IRS filing, where it previously aimed to build “safely benefits humanity” for general AI. It is now focused on “ensuring that general AI benefits all of humanity.”

RAND Corporation senior policy researcher Edward Geist said that the advanced AI envisioned by early AI safety advocates is fundamentally different from current large language models. The change in terminology also reflects companies’ desire to signal to investors and policymakers that they will not retreat from economic competition due to safety concerns.

Recently, Anthropic completed a $30 billion funding round, with an estimated valuation of about $380 billion; OpenAI is pushing a funding round supported by Amazon, Microsoft, and Nvidia, potentially reaching up to $100 billion. Meanwhile, Anthropic has publicly disagreed with the U.S. Department of Defense over full access to Claude for the Pentagon, leading to uncertainty over its defense contracts. (Decrypt)