Anthropic Rolls Out Election Safeguards for Claude AI Ahead of US Midterms

Anthropic, the artificial intelligence company behind the Claude chatbot, announced on Friday a series of new election integrity measures designed to prevent AI from being weaponized to spread disinformation or manipulate voters in the run-up to the 2026 US midterm elections and other major contests around the world this year.

The San Francisco-based company has crafted a multi-pronged approach that includes automated detection systems, stress testing against influence operations and a partnership with a nonpartisan voter organization – measures that reflect growing pressure on AI developers to monitor how their tools are used during election seasons.

Anthropic’s usage policy prohibits Claude from being used to conduct deceptive political campaigns, generate false digital content intended to influence political discourse, commit voter fraud, disrupt voting infrastructure, or disseminate misleading information about voting processes.

To enforce these rules, the company said it has put its latest models through a series of tests. Using 600 prompts (300 malicious requests combined with 300 legitimate requests), Anthropic measured how reliably Claude fulfilled appropriate requests and denied problematic requests. Claude Opus 4.7 and Claude Sonnet 4.6 responded adequately 100% and 99.8% of the time, respectively.

The company also tested its models against more advanced manipulation tactics. Using multi-turn simulated conversations designed to mirror the step-by-step methods bad actors might use, Sonnet 4.6 and Opus 4.7 responded appropriately 90% and 94% of the time when tested against influence operations scenarios.

Anthropic also tested whether its models could autonomously perform influence operations: planning and executing a multi-step campaign, from start to finish, without human assistance. With the necessary safety measures in place, the latest models refused to perform almost any task, the company said.

On the issue of political neutrality, the company conducts assessments before each model launch to measure how consistently and impartially Claude handles cues that express views across the political spectrum. Opus 4.7 and Sonnet 4.6 scored 95% and 96% respectively.

For users looking for voting information, Claude will display an election banner that directs them to TurboVote, a nonpartisan source from Democracy Works that provides reliable, real-time information on voter registration, voting locations, election dates, and voting data. A similar banner is planned for Brazil’s elections later this year.

Anthropic said it plans to continue monitoring its systems and refining its defenses as the election cycle progresses. Declutter contacted Anthropic for comment on the findings but did not immediately receive a response.