OpenAI provides its cybersecurity model to the EU for evaluation, while Anthropic withholds its advanced Mythos tool amid regulatory scrutiny.
OpenAI has begun sharing its new cybersecurity model with the European Union for security assessments, addressing concerns over cyberattack risks on critical software. The move aims to align with EU safety priorities and democratize access to defensive AI tools for trusted institutions.
Anthropic, however, has not released its Mythos model in the EU, citing ongoing discussions and heightened scrutiny due to its advanced capabilities in identifying security vulnerabilities. The EU seeks broader access to such tools to bolster cyber defenses across member states.
The differing approaches highlight varying levels of cooperation among AI developers with EU regulators, as the bloc pushes for stricter oversight of high-risk AI applications.