A recent study by Dutch researchers, led by Joost van der Burgt from the Dutch Data Protection Authority (DPA), highlights significant biases in AI chatbots when asked for voting advice. As the Netherlands approaches its election on October 29, the DPA aimed to understand how these technologies might influence voter decisions. The researchers tested popular chatbots, including ChatGPT and Gemini, by creating diverse fake voter profiles and querying them about party recommendations. Surprisingly, 56% of responses directed users towards the far-right Freedom Party or the Labour-Green Left coalition, oversimplifying the country’s rich political landscape. Van der Burgt cautioned against relying on such biased tools and emphasized the importance of traditional voting aids and informed decision-making. He stressed that democratic participation is vital and urged voters to align their choices with their values while avoiding AI-generated political guidance.
Source link
Share
Read more