Dutch Data Protection Authority Warns: AI Chatbots Provide Biased Voting Advice
den haag, woensdag, 22 oktober 2025.
The Dutch Data Protection Authority (AP) warns voters against using AI chatbots for voting advice. According to research, these bots provide unreliable and often biased advice, with an excessive focus on the PVV and GreenLeft-Labour Party. This can undermine the integrity of elections. The AP advises voters to use traditional voting aids such as Kieskompas and StemWijzer instead.
Dutch Data Protection Authority Warns: AI Chatbots Provide Biased Voting Advice
The Dutch Data Protection Authority (AP) expresses significant concerns about the use of AI chatbots for voting advice. According to recent research, these recommendations are often unreliable and clearly biased, which can damage the integrity of elections. The AP advises voters to use traditional voting aids such as Kieskompas and StemWijzer instead [1][2].
Research Methods and Results
In the study, where voting advice was requested a total of 21,000 times, the regulator compared four well-known chatbots — those from ChatGPT, Gemini, Mistral, and Grok — with established voting aids like Kieskompas and StemWijzer. Regardless of the user’s question or request, the AI bots frequently recommend the PVV or GreenLeft-Labour Party. In 56 percent of cases, these parties come out on top, and with one chatbot, it is even more than 80 percent of the time [1][2][3].
Parties Rarely Recommended
Parties such as D66, SP, VVD, and Party for the Animals are much less frequently recommended as the first choice. Parties like BBB, CDA, SGP, and DENK are almost never given as primary voting advice, even when the user reflects their positions exactly in conversations with the bots [1][2][3].
Ethical and Privacy Issues
The use of AI bots as a supplement to traditional voting aids is increasing, raising new ethical and privacy issues. The advice from chatbots is based on unverifiable data used to train them and information from the internet, which can be incorrect or outdated. This can lead to a distorted view of the political landscape and influence voters with inaccurate information [1][2][3].
Reactions and Recommendations
Monique Verdier, vice-chair of the AP, states: ‘Chatbots may seem like smart helpers, but as voting aids, they consistently miss the mark. As a result, voters may unknowingly receive advice to vote for a party that does not best align with their preferences. This directly impacts a cornerstone of democracy: the integrity of free and fair elections.’ The AP calls on chatbot providers to prevent their systems from being used for voting advice [1][2][3].
Implications for Media Literacy and Democracy
The use of AI in the dissemination of information and voting advice has far-reaching implications for media literacy and democracy. It is crucial that voters are critical and use reliable sources. Traditional voting aids offer more transparency and are developed to provide impartial and accurate advice [1][2][3].
Practical Tips for Readers
To identify fake news and biased voting advice, readers can follow these tips: 1. Always use multiple reliable sources for information. 2. Check the background and reputation of the information sources. 3. Be critical of extremely emotional or sensational messages. 4. Consult traditional voting aids such as Kieskompas and StemWijzer for impartial advice. 5. Stay informed about the latest developments in politics and media [1][2][3].