Anthropic changes usage policy to ban model abuse and election interference

Anthropic has revised the usage policy governing its Claude models, according to TechCrunch. The updated policy explicitly prohibits repeated abuse of the model in extreme cases, while permitting ordinary user frustration and criticism to continue unrestricted.

The revised rules also add or clarify prohibitions in several areas. Election interference and deceptive campaigns are now barred under the policy. The policy further addresses weapons software and surveillance applications, placing restrictions on these uses as well.

TechCrunch reports the changes as part of Anthropic’s ongoing policy for Claude, without further elaboration on the company’s internal deliberations or enforcement mechanisms. The published summary does not specify what prompted this particular round of revisions, nor does it detail how violations will be identified or penalized.

The policy update covers four distinct areas, model abuse, election interference, weapons software, and surveillance, each addressed as a separate category within the revised terms. No additional categories or exceptions beyond these four were described in the reporting.

This bulletin reflects only what TechCrunch’s account makes available. The source material does not include the full text of the policy itself, statements from Anthropic representatives, or reaction from outside parties. Readers seeking the complete policy language or company commentary should consult the original reporting directly.