Research consistently shows that ChatGPT, Claude, Gemini, and Grok are unreliable sources for election news, with citation failure rates above 60% and false claim rates of 35%. With the 2026 US midterms 167 days away, a growing cohort of voters is expected to use chatbots as their primary news source. Studies show models systematically misattribute quotes, fabricate links, and cannot distinguish credible journalism from disinformation. Publisher licensing deals have not improved accuracy. Labs have positioned chatbots as auxiliary sources, but this advice is irrelevant for voters who rely on them as primary sources. Neither the US regulatory framework nor the labs' own infrastructure is equipped to address provenance or misinformation at scale before November.
