Google and meta are restricting political ads in the EU. But what if the next election battlefield isn’t social media… it’s inside AI answers? Studies show ChatGPT leans left on key issues, and 40%+ of its political profiles contain factual errors. If voters start turning to AI for “neutral” info, what’s stopping campaigns from quietly optimizing prompts to trigger subtle bias? If campaigns can optimize prompts to trigger bias, they can shape how voters see the world, without running a single ad. Today, brands are already competing to control how they show up in LLMs. Tomorrow… could political parties do the same?