ASKED FOR POLLS. GOT BETS.

The real news

The story, straight

A poll asks people how they plan to vote. A prediction market lets people bet on who will win.

Both can discuss the same election. They are not measuring the same thing.

So naturally, the chatbot arrives at democracy's help desk holding a betting slip. Very reassuring.

Does the constitution come with casino points?

NPR reports that an ongoing study by Innovating for the Public Good, a democracy-research nonprofit, tested 7 chatbot services on authentic voter questions.

All the platforms tested commonly cited prediction markets such as Polymarket and Kalshi when asked who was winning and what the latest polls showed. That is not the same as asking voters.

The trouble also reached the practical business of voting. In the study's October 2026 data, 13.9% of answers to voting-logistics questions were outdated or inaccurate.

Errors in 5.9% could cause a missed deadline, a visit on the wrong day or a skipped required step. Those percentages describe answers, not voters, and do not apply to every election question.

This does not mean the chatbots told people to gamble. It means they commonly supplied betting-market information in response to polling questions.

The distinction matters: a confident answer can still be the wrong kind of answer.

The useful move: check voting rules directly with your official election office, and check whether a supposed poll actually surveyed voters. Democracy already has enough paperwork without adding a betting slip.

Repost so the voter guide stays out of the casino.

Thanks for supporting Lil'meme. Your support is appreciated.