67% of AI Chatbot Election Answers Incomplete: Bipartisan Letter to 4 Agencies Seeks Regulation

The July 9, 2026 letter from two House members to four federal agencies is not just a political statement—it is a legal stress test for the regulatory frameworks governing artificial intelligence and elections. With a study showing that 67% of AI chatbot responses about elections are incomplete, the letter directly challenges whether current laws give DHS, DOJ, CISA, or the FEC the tools to compel accuracy. For legal and regtech professionals, the scramble to answer these questions before November could reshape compliance obligations for AI developers and election technology vendors overnight.

In an unusual bipartisan move, Representatives Mike Lawler (R-NY) and Josh Gottheimer (D-NJ) have pressed four federal agencies to immediately coordinate on countering the threat of artificial intelligence to the integrity of the 2026 midterm elections. In a letter sent on July 9, 2026 to the Department of Homeland Security, the Department of Justice, the Cybersecurity and Infrastructure Security Agency, and the Federal Election Commission, the lawmakers explicitly tied the accuracy and neutrality of AI-driven tools to the health of American democracy, warning that millions of voters now rely on chatbots to research candidates and ballot measures. The letter, first reported by The Hill, represents one of the first concrete legislative efforts to force a government-wide risk assessment of AI as an election interference vector, moving beyond general warnings about deepfakes to focus on the everyday, consumer-grade chatbots that shape voter knowledge.

“

With a study showing that 67% of AI chatbot responses about elections are incomplete, the letter directly challenges whether current laws give DHS, DOJ, CISA, or the FEC the tools to compel accuracy.

Central to the lawmakers' argument is a study conducted approximately two years earlier, in 2024, which found that over two-thirds of AI chatbot responses to election-related queries were incomplete, and all models performed poorly when tested on voting information. The study, while not named in the letter, underscored a fundamental mismatch: voters increasingly treat AI chatbots as trustworthy digital librarians, even as the technology's training on the open, often misinformation-filled internet leaves it ill-equipped to deliver factual, unbiased civic guidance. The letter quotes the lawmakers directly: 'These sources present an environment filled with unverified information, where anyone with an opinion can shape the narrative. By relying on this information, systems cannot reasonably present a full picture of unbiased information to voters.' The framing is significant because it moves the debate from content moderation—a politically charged arena—to a consumer-protection and national-security footing, one that both parties can rally around.

The four agencies now under pressure each possess distinct but overlapping authorities that, if combined, could create a formidable wall against AI election interference. DHS, through its Office of Intelligence and Analysis and its designation of election infrastructure as critical infrastructure, can disseminate threat intelligence to state and local officials. CISA, as the nation's cyber defense agency, has deep relationships with election technology vendors and can issue binding operational directives. The DOJ can investigate and prosecute foreign or domestic actors who weaponize AI to suppress votes or spread disinformation. And the FEC, though traditionally focused on campaign finance, is increasingly asked to rule on whether AI-generated content constitutes an in-kind contribution or requires disclaimers. The letter's call for an interagency task force reflects a recognition that no single agency has the mandate to address the problem end-to-end; a coordinated response, however, could bridge gaps that adversaries are already exploiting.

Legally, the push opens a new chapter in election law and AI governance. It raises immediate questions about agency authority: can CISA compel AI companies to adopt accuracy standards? Does Section 230 of the Communications Decency Act shield chatbot providers from liability for election disinformation? Would an FEC rule requiring labeling of AI-generated political content survive First Amendment challenges? The bipartisan sponsorship signals that Congress may be ready to legislate if the agencies demur, potentially reviving stalled bills like the AI Election Integrity Act or the Protect Elections from Deceptive AI Act. Yet with the 2026 general election mere months away, the timeline is exceedingly tight; any agency rulemaking would almost certainly face legal challenges that push implementation past November.