ChatGPT and other AI bots made huge errors before Scottish election, study finds
TL;DR
The UK Electoral Commission has called for new legal controls over misinformation from AI chatbots, after thinktank Demos found they had made serious mistakes during the recent Scottish election. Demos said its investigation had found that AI services gave voters misinformation to 34% of the questions it posed, raising worrying questions about the lack of regulation of AI platforms in the UK. Tools made up fake scandals, invented candidates or gave the wrong date.
Nauti's Take
Opportunity in the study: it delivers hard numbers — 34% misinformation — finally giving regulators evidence to build serious AI rules for democracies. Risk signal is clear: chatbots inventing scandals, hallucinating candidates or spitting out wrong dates ahead of elections is a democratic security problem.
Practically, voter information should not be delegated to ChatGPT but pulled from official sources. Platforms need binding election guardrails, not voluntary pledges.
Sources