AI chatbots have been shown to be “inaccurate and evidently prejudiced” when offering voting advice, the Netherlands’ data watchdog has reported, raising alarms about a potential threat to democracy just eight days before the country’s general election.
Testing Reveals Repeated Partiality in Chatbot Recommendations
The quartet of AI assistants evaluated by the AP “frequently recommend identical two parties”, irrespective of the voter’s inquiry or input, the watchdog explained in a report released ahead of the late October vote.
For over 50% of instances, the chatbot suggested either the far-right Freedom party (PVV) of the politician Geert Wilders or the progressive GroenLinks-PvdA headed by the ex-EU commissioner Frans Timmermans.
Certain groups, such as the centre-right CDA, “are rarely ever cited”, even when the user’s input precisely aligns with the stances of these parties, the study pointed out.
Official Warns Against Using AI Chatbots for Voting Guidance
The vice-chair of the AP, Monique Verdier, stated that while chatbots may seem like clever tools, “as a voting aid, they consistently fail”. Voters were being pushed toward a party that might not correspond to their ideological beliefs, she emphasized.
“This directly impacts a cornerstone of democracy: the honesty of open and just votes,” said Verdier. “We therefore warn against using AI chatbots for election recommendations, as their functioning is opaque and hard to confirm.”
Voting Background and Party Dynamics
The Dutch head to the polls on October 29 in an election that will be closely watched across the continent for the performance of the PVV.
The far-right party of Geert Wilders is leading the polls but the lead over the GroenLinks-PvdA and CDA appears to be narrowing, with numerous electors still undecided.
All major parties have rejected a coalition with the Freedom party, implying the runner-up party is most likely to supply the future head of government.
Inadvertent Prejudice Linked to AI Operation
The authority emphasised that the AI systems were not intentionally prejudiced, and that their “noted flaws” are a consequence of the manner in which AI assistants function.
