The AI Election Whisperer: Why Chatbots Can't Be Trusted with Your Vote
There’s something deeply unsettling about the idea of an AI chatbot telling you how to vote. It’s like asking a stranger on the street to decide your future—except this stranger is a machine, and it’s not even sure what it’s talking about. A recent study by the civil liberties group Liberties has exposed just how unreliable AI chatbots are when it comes to political advice, and it’s a wake-up call we can’t ignore.
The Illusion of Authority
One thing that immediately stands out is how convincingly authoritative these chatbots sound. They don’t hesitate, they don’t stutter—they deliver their recommendations with the confidence of a seasoned political analyst. But here’s the catch: they’re often dead wrong. In the study, ChatGPT and Gemini, two of the most popular AI models, routinely misclassified voter profiles, omitted relevant parties, and even suggested parties that weren’t running. What makes this particularly fascinating is how easily users might be swayed by this faux confidence. If you take a step back and think about it, this isn’t just a technical glitch—it’s a trust issue.
The Tisza Paradox
A detail that I find especially interesting is the chatbots’ bizarre treatment of Hungary’s Tisza party. Despite Tisza’s landslide victory in the 2026 parliamentary elections, ChatGPT failed to recommend it to Tisza-aligned voters a staggering 90% of the time. Instead, it often suggested smaller, irrelevant parties. What this really suggests is that AI systems struggle with dynamic political landscapes. Tisza’s rapid rise to prominence post-2024 wasn’t on the radar of these static training models. From my perspective, this isn’t just a failure of data—it’s a failure of adaptability. Politics moves fast, and AI, it seems, is stuck in the slow lane.
The Fidesz Advantage
On the flip side, Fidesz, the long-standing ruling party, was consistently recognized and recommended by the chatbots. This raises a deeper question: are AI systems inadvertently favoring established parties? It’s not hard to see why this could be problematic. In a more competitive election, such bias could tip the scales unfairly. What many people don’t realize is that AI’s ‘neutrality’ is a myth. It’s only as neutral as the data it’s trained on, and if that data is skewed toward historical dominance, the results will be too.
The Regulatory Black Hole
Here’s where things get really concerning: AI chatbots fall into a regulatory no-man’s-land. The EU’s AI Act and Digital Services Act touch on systemic risks to elections, but they don’t explicitly cover chatbots providing political advice. Personally, I think this is a glaring oversight. If AI is going to play a role in democracy—even an unintentional one—it needs to be held accountable. Transparency, accuracy, and consistency should be non-negotiable. Yet, as the study shows, these systems are opaque, unstable, and often wrong.
The Human Factor
What this debate often misses is the human element. Voting isn’t just about matching policies to preferences—it’s about values, emotions, and context. AI can’t replicate that. In my opinion, the idea that a chatbot could ever truly understand the nuances of a voter’s decision-making process is absurd. It’s like trying to solve a moral dilemma with a calculator.
Looking Ahead
So, where do we go from here? For starters, AI providers need to stop pretending their systems are ready for this level of responsibility. Until they can guarantee transparency and accuracy, they shouldn’t be offering political advice. But this is also a moment for voters to step back and ask themselves: do we really want machines meddling in our democracy? If you ask me, the answer is a resounding no.
Final Thoughts
The Liberties study isn’t just a critique of AI—it’s a reminder of what’s at stake. Democracy is fragile, and it relies on informed, thoughtful participation. AI chatbots, with their inconsistencies and biases, are the opposite of that. As we move forward, we need to be vigilant about where and how we draw the line. Because if we don’t, we risk handing over the keys to our democracy to systems that don’t even understand what they’re doing. And that, in my opinion, is a future we can’t afford.
