AI Chatbots and Voting Advice: A Distorted Picture
Artificial intelligence may appear neutral, but that’s often just an illusion. Especially during election season.
The Dutch Data Protection Authority (Autoriteit Persoonsgegevens) warns that AI chatbots, including popular generative AI tools, may provide biased or inaccurate voting advice. Such advice is often based on incomplete or skewed data, meaning voters could be influenced by misleading information, a serious risk in the digital age.
A Fundamental Problem
AI systems have no moral compass. They generate answers based on training data, algorithmic patterns, and user input.
What’s often missing is transparency (as required under Article 13 of the GDPR) and human oversight.
The AI Act therefore imposes strict requirements on general-purpose AI systems used at scale, including transparency obligations, fundamental rights impact assessments (FRIA), and human supervision.
For organizations deploying AI, even when using third-party tools, it’s essential to understand the technology and its risks. This is especially true when AI influences public opinion, elections, or social decision-making. In such contexts, the legal and ethical bar is set high.
Supporting Responsible AI Governance
At Data & Business Legal, I help organizations navigate these challenges through the role of the Certified AI Compliance Officer (CAICO®).
This includes impact assessments, AI governance frameworks, contractual safeguards, and policies for responsible use.
The DPA’s warning is not just for tech companies: it applies to every organization using AI tools.
After all, algorithms have no conscience, but their users do.
Does your organization use AI responsibly?
Is there clear policy, oversight, and awareness in place?
Perhaps it’s time for an AI audit or a tailored compliance strategy, I’m happy to help you take the next step.
