When politicians use AI to write speeches, voters trust them less

Dutch Prime Minister Jetten and other politicians are increasingly turning to generative AI — software that can write text by learning patterns from huge amounts of existing writing — to help craft their political messages and speeches.

But a new study reveals an uncomfortable truth: when voters find out that AI was involved in writing a politician's statement, they immediately trust that politician less. The research shows this trust drop happens even when the actual message is exactly the same, word-for-word.

This raises a real question for modern politics: what does it mean to be authentic as a leader when machines help shape your public voice? If a message is identical whether a human wrote it alone or with AI assistance, why does knowing about the AI matter so much to voters? The answer seems to be that people care not just about what is said, but who they believe truly said it.

For politicians, the finding is a warning. Using AI tools might save time, but if voters discover it, the cost to their credibility could be steep.