Anthropic Updates 2026 Usage Policy, Tightens Weapons Ban

The notable shift here is not the existence of a weapons ban but its move from a general principle to an explicit prohibition, alongside tighter limits on surveillance and tracking — the two areas where AI governance debates have been most contested. This matters most for enterprise and developer customers whose compliance teams treat usage policies as contractual risk, not marketing language, and who may need to re-examine deployments touching monitoring or defense-adjacent work. Whether other major model providers match this specificity, or keep looser language, is the open question worth watching.