Anthropic to lift blanket ban on personalized voter targeting
From 12 November, Anthropic is removing the blanket ban on personalized targeting for elections and campaigns from the Usage Policy. Deceiving voters and misusing their personal data remain prohibited. According to the company, the ban on cruel treatment of models does not cover ordinary frustration or research.
Anthropic has published an update to the Usage Policy that will take effect on 12 November. It removes the blanket ban on personalized targeting for elections and campaigns. According to the company, the ban also affected legitimate civic activities, such as preparing information for voters in other languages. Targeting based on deception or misuse of voters’ personal data remains prohibited.
A new section consolidates the existing restrictions on deceptive political and commercial campaigns. It covers concealing the source of messages, artificially amplifying content through fake accounts or posts, and building tools and infrastructure for such campaigns. The election rules prohibit deceiving voters and disrupting elections, for example by providing false information about candidates or how to vote, impersonating candidates or election officials, and suppressing voter turnout. The TechCrunch article describes the changes as new bans; Anthropic says restrictions on deceptive campaigns were already in place and are now being consolidated.
The update also explicitly prohibits sustained and gratuitous cruel or abusive treatment of Claude models. According to the company, this covers only extreme cases, rather than ordinary frustration, disagreement, dark creative themes, or testing and research. According to the TechCrunch article, Claude models have been trained since the August update to end persistently harmful or abusive conversations. Ending the conversation remains the main mechanism for enforcing this ban, according to the article in The Verge.
Why it matters
Assessments of election-related use will hinge on how targeting is carried out and how data is handled, rather than personalization itself. For deceptive campaigns, the rules explicitly cover the development of tools and infrastructure as well, so they also apply to providers of these solutions.
What was added since the original report
Verified updates
-
The ban on cruel treatment does not apply to ordinary frustration, disagreement, dark creative themes, or testing and research.; Since the August update, Claude has been trained to end persistently harmful or abusive conversations.
- The ban on cruel treatment does not apply to ordinary frustration, disagreement, dark creative themes, or testing and research.
- Since the August update, Claude has been trained to end persistently harmful or abusive conversations.
-
The updated rules take effect on 12 November 2026.; Anthropic is removing the blanket ban on personalized targeting for elections and campaigns.; Deceptive targeting and misuse of voters’ personal data remain prohibited.
- The updated rules take effect on 12 November 2026.
- Anthropic is removing the blanket ban on personalized targeting for elections and campaigns.
- Deceptive targeting and misuse of voters’ personal data remain prohibited.
Two audiences, two different impacts
What this means
For individuals
If you are preparing information for voters, the removal of the blanket ban from 12 November opens up opportunities for legitimate personalized use of Claude models, such as messages in different languages.
For a business
Companies using Claude models in marketing must also include the source of messages, fake accounts and infrastructure for artificially amplifying content in their compliance checks. The consolidated ban also applies to commercial campaigns.
Risks and complianceCheck the original
Event sources
confirmed by 2 independent sources · 3 publishers, 2 independent. We count feeds from the same owner only once.