Google, OpenAI, and Anthropic Plan Self-Regulatory AI Safety Group

Google, OpenAI, and Anthropic are reportedly collaborating to establish an industry self-regulatory body named the Standards Authority for Frontier AI (SAFA), targeted to launch by late 2024 or early 2025. The independent organization aims to set safety standards and testing protocols for cutting-edge AI models without direct government oversight. As government regulation of artificial intelligence lags behind rapid technical advancements, the top AI labs are proactively establishing their own auditing mechanisms. This initiative could define global industry standards for model testing and influence future public policy and safety legislation. SAFA plans to support third-party model evaluation before deployment, standardize safety incident reporting, and define qualifications for independent auditors. The founding members are also debating whether SAFA should conduct safety testing directly due to concerns that government bodies, such as the U.S. AI Safety Institute, lack adequate resources.