White House Moves to Give US Agencies Anthropic Mythos Access

The White House is granting federal agencies access to Anthropic’s advanced AI model, Mythos, despite cybersecurity concerns outlined in an internal memo obtained by Bloomberg. Major US agencies will receive a tailored version of Mythos, a cutting-edge AI system developed by Anthropic PBC, marking a significant expansion of government AI adoption. The initiative, announced in April 2026, aims to enhance agency capabilities but raises alarms over potential vulnerabilities in critical infrastructure and data protection. Cybersecurity experts warn the model’s deployment could escalate risks, including adversarial attacks or unintended data leaks, due to its unprecedented computational power. The move reflects the government’s push to integrate next-gen AI while balancing innovation with national security, setting a precedent for future tech deployments.