OpenAI outlines responsible AI efforts across Europe

OpenAI said it is expanding its responsible AI work across Europe as the EU AI Act enters a new phase. The company said its approach focuses on safety, security, transparency, and provenance, and that it aims to align its practices with the EU framework while supporting innovation. The post describes several measures OpenAI says it has used for years, including prerelease model testing, published system cards, outside expert redteaming, and the public Model Spec. It also points to internal governance frameworks that are meant to identify and manage serious risks from advanced AI systems. OpenAI said it is also working on transparency for AIgenerated content through provenance tools such as Content Credentials and SynthID watermarks, with plans to expand those efforts to audio and other formats. In cybersecurity, the company said it has been working with European agencies, privatesector partners, and infrastructure operators to support defensive use cases through its Trusted Access for Cyber program and related initiatives.