ChatGPT and other AI chatbots provided false information to voters in 34% of election-related questions during Scotland's recent vote, prompting the Electoral Commission to call for new legal controls on AI-generated misinformation.
A study by the thinktank Demos uncovered serious errors in how major AI services handled Scottish election queries, with tools fabricating scandals, inventing candidates, and providing incorrect dates.
The Electoral Commission's response signals growing concern about AI's role in political discourse. The regulator is now pushing for new legal frameworks to govern how AI chatbots handle election-related content and misinformation.
Demos tested various AI services with questions about the Scottish election and found alarming failure rates. The chatbots didn't simply refuse to answer or express uncertainty—they generated plausible-sounding but entirely false information. This included invented political scandals and candidates that never existed.
The findings highlight a critical vulnerability in the information ecosystem during elections. Voters increasingly turn to AI tools for quick answers, yet these systems can confidently deliver false narratives without adequate safeguards.
The timing is significant. As general elections approach, regulators and tech companies face pressure to address misinformation risks. Current approaches rely largely on voluntary measures and internal safety protocols from AI developers, which have proven insufficient.
The Electoral Commission's call for new controls suggests potential regulatory approaches could include mandatory fact-checking systems, clearer disclaimers about AI limitations, or restrictions on generating election-specific content during voting periods.
The Demos findings will likely influence broader policy discussions about AI governance in the UK and beyond. Other democracies grappling with election integrity face similar challenges as AI capabilities expand.
Neither ChatGPT nor other AI services tested have announced specific changes in response to the study, though the Electoral Commission's formal stance adds regulatory pressure to the issue.
Uber's weekly AI agent requests have grown nearly tenfold since February, yet the company has held spending flat since April after exhausting its entire 2026 AI budget in Q1.
A recent paper shows artificial intelligence often diagnoses and treats patients better than human physicians. The findings are prompting difficult conversations within the medical community about the profession's evolving role.
The Relay Q, launching next year, represents the latest push to establish voice as the primary interface for human-computer interaction, challenging the keyboard's decades-long dominance.
An Anthropic researcher demonstrated automated systems that can identify and correct misaligned behaviors without compromising overall performance. The systems improved on all 10 tested benchmarks measuring specific problematic outputs.