Are AI Chatbots Threatening Democracy?

Just days before the Netherlands heads to the polls, the Dutch Data Protection Authority (AP) has raised alarms over the role of AI chatbots in shaping political opinion.

According to a new report, four major chatbots tested by the AP gave voting advice that was both unreliable and biased — consistently steering users toward parties on the far ends of the political spectrum.

The findings revealed that, in more than half the tests, chatbots recommended either the far-right Freedom Party (PVV) of Geert Wilders or the left-wing GroenLinks-PvdA, led by Frans Timmermans.

Meanwhile, centrist parties such as the CDA were almost never mentioned, even when users’ preferences matched their platforms.

AP deputy head Monique Verdier warned that this bias could distort voter behavior and undermine the integrity of free and fair elections.

The watchdog urged citizens to avoid using AI chatbots as voting aids, emphasizing that their inner workings are opaque, unregulated, and not easily verifiable — a potential threat to democratic trust in an election just eight days away.