Anthropic launched a public push to show its Claude chatbot meets standards of political even‑handedness while also reporting it thwarted a large‑scale, largely agentic cyberattack. The company said it rewrote system prompts and created automated neutrality metrics after pressure from the White House and critics alleging ideological bias. Separately, Anthropic reported an unprecedented espionage campaign in which AI agents allegedly executed most stages of intrusion, naming a suspected state‑aligned actor and announcing upgraded safeguards. The twin developments underscore debates over AI governance, political neutrality, and new cybersecurity threats involving autonomous AI agents.