Anthropic and OpenAI Revise Safety Policies Amid AI Competition
- Anthropic has removed a key safety pledge from its Responsible Scaling Policy, no longer promising to stop AI training if safeguards are not in place.
- OpenAI revised its mission statement, removing the word “safely” and focusing on ensuring AI benefits all of humanity.
- Both companies are adjusting their policies as they seek to strengthen their commercial positions, with Anthropic valued at $380 billion and OpenAI’s funding round potentially reaching $100 billion.
- Anthropic is in a dispute with the U.S. Department of Defense over access to its Claude AI system, unlike other major labs like Google and xAI.
These policy shifts by Anthropic and OpenAI reflect a strategic pivot in response to intense competition in the AI industry and changing political dynamics. The removal of specific safety language signals a focus on economic competition rather than unilateral commitments to safety.
Anthropic’s decision not to halt training without full safeguards marks a significant change in its approach, aligning more closely with industry trends prioritizing rapid development over previous safety pledges. (Source)
