Official announcement from the US AI Safety Institute or Anthropic

OpenAI and Anthropic Partner with US for AI Safety Testing

Aug 29, 2024, 10:39 AM

The United States has announced that OpenAI and Anthropic have agreed to provide early access to their major new AI models to the US AI Safety Institute. This agreement aims to facilitate the evaluation of the capabilities and risks associated with these AI models. The collaboration is part of efforts to advance AI safety research, testing, and evaluation. The US government agency will test these models before their public release, marking a significant step in AI safety and regulation.