A tech trade association that counts Anthropic and OpenAI as members is pushing lawmakers to require that some high-risk AI models publish their approach to safety evaluations.
TechNet is also telling Congress that all publicly available models should offer transparency of their intended uses and limitations.