What happens when millions ask a chatbot who to vote for?
Here is a habit that crept up on us. Millions of Americans now turn to AI chatbots for election information. NPR reports that the bots get voting facts wrong, and that they also recommend candidates. This lands the day before the US midterms.
Think about what a chatbot is to many people. Not a search page with ten blue links to compare. One calm, confident voice that answers in full sentences and never says "I don't know" the way a person would.
Is this about lazy users?
It is tempting to say people are just lazy and easy to fool. But is that really the problem? Or is it that the product sounds certain even when it is wrong?
The second one. A wrong date for when polls close, delivered in a confident tone, does damage in a way a messy search result does not. The flaw is built into how the product talks, not into the person asking.
And here is what nobody teaches. At school you learn to check a source, to ask who wrote something and why. Almost no one learns to check an answer from a chatbot, or to notice when it nudges toward a particular candidate while sounding perfectly neutral.
There is a lot we still do not have solid data on, like how much these answers actually change how people vote. What we can say is simple: a tool that millions treat as a neutral expert should be held to the standard of one, especially on the one topic where a mistake cannot be undone after election day.