AI Chatbots Provide Inaccurate Voting Information
The New York Attorney General tested multiple AI chatbots and found they frequently provided inaccurate voting information, including false registration deadlines and incorrect polling locations. The office issued a consumer alert warning voters not to rely on chatbots for election information and to use only official government sources instead.
What happened
The Office of the New York Attorney General tested multiple AI-powered chatbots by posing sample questions about voting and found that they frequently provided inaccurate information in response.
Specific errors included chatbots providing false information about voter registration deadlines and early voting locations. When asked about early voting in Kings County, one chatbot mistakenly advised voters could vote at any designated early voting site within the county, when in fact those in New York City, including all of Kings County, are assigned specific poll sites based on residence.
Reported impact
New Yorkers relying on chatbots for voting information risk being misinformed and could lose their opportunity to vote due to inaccurate information, including missing registration deadlines or attempting to vote at incorrect locations.
Index analysis
Large language models routinely produce inaccurate and unreliable information. In a July study, the Center for Democracy & Technology found that in response to 77 different election-related queries, more than one-third of answers generated by AI chatbots included hallucinations about polling locations, voting accessibility, and permissible voting methods.
Governance controls
The New York Attorney General's office issued a consumer alert on November 1, 2024, warning voters against relying on AI chatbots for election-related questions and urging them to use only official government sources for voting information.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.