European institutions are seeking access to high-powered AI models to protect them from cyberattacks after Anthropic denied EU regulators access to its powerful Mythos model.
French AI startup Mistral is reportedly in talks with European banks to provide a cybersecurity-focused model, similar to Mythos, while Anthropic rival OpenAI said it would grant the EU access to its latest model.
Governments and businesses are racing to detect and fix software vulnerabilities before AI systems get to them first, with access to the tech intertwining with US foreign policy: AI safety is on the agenda at the Trump-Xi summit, while three Japanese banks were reportedly given access to Mythos following the US Treasury secretary’s meeting with Japan’s prime minister.