Report: Voters Use ChatGPT and Other Chatbots for Midterm Ballot Research, Raising Bias and Privacy Concerns

OpenAI's report that voters are using ChatGPT for midterm ballot research, with users like Adam Johnson relying on the chatbot to parse candidate information, highlights a surge in political‑focused AI traffic to public websites. For website owners, this means a spike in bot‑generated queries to election‑related pages, policy pages, and candidate bios. Because ChatGPT often scrapes full‑text content to feed its model, sites may see higher-than‑usual request volumes, potential bias amplification if the bot preferentially extracts certain language, and increased privacy risk as user‑level data (e.g., IP addresses) could be inadvertently logged by the AI service. **Actionable tips:** 1. Update your llms.txt to explicitly disallow OpenAI’s crawler (User‑Agent: ChatGPT) from accessing sensitive political pages, using `Disallow: /election/` or similar paths. 2. Deploy a bot‑detection rule in your analytics stack that flags high‑frequency, low‑session‑duration visits from known AI IP ranges (e.g., 13.107.*) and redirects them to a static summary page. 3. Implement a consent banner that informs visitors that AI bots may be scraping content and offers an opt‑out, reducing privacy exposure and helping you comply with emerging data‑use guidelines.