AI Labs Seek Shared Safety Protocols Amid Rising Model Capabilities

Three major AI labs - OpenAI, Anthropic, and Google DeepMind - are engaging in discussions to establish shared industry standards for AI model safety protocols. The CEOs of these companies have publicly called for US-led collaboration on AI safety standards, including international forums for testing and risk analysis.

The proposed core idea is to establish shared protocols for testing frontier AI models before they're released to the public, including independent evaluations, pre-release safety reviews, and standardized risk assessments. However, the labs hold different views on how much government involvement is appropriate versus industry self-regulation.

Anthropic has leaned more explicitly toward government partnership, while OpenAI emphasizes its commitment to advancing voluntary AI standards regardless of government intervention. Google DeepMind's internal safety research agenda remains somewhat less public compared to its peers.

The Frontier Model Forum, a consortium including these companies, was created to address the safety of powerful AI models. Recent incidents involving AI models have sharpened the urgency of these discussions.