UK Electoral Commission is pushing for AI regulations (chatbots got 34% of election answers wrong)

2026-05-21 · Source: Artificial Intelligence · Field: Government & Public Sector — Public Policy & Governance, Regulatory & Compliance · Depth: Fundamental Awareness, quick

Summary

The UK Electoral Commission's head, Vijay Rangarajan, advocates for new legal restrictions on AI chatbots following a Demos think tank report on misinformation during Scottish elections. The study tested five free AI tools with 75 election questions, revealing an average 34% error rate. Replika performed worst at 56%, ChatGPT at 46%, and Google Gemini at 22%, while Grok had the lowest error rate at 9%. Chatbots generated fake candidates, invented scandals, and provided incorrect election dates and voting rules, often lacking official sources. This is significant as 20% of UK voters, approximately 10 million people, used AI platforms for election information. Currently, the UK lacks laws holding AI companies accountable for such misinformation, prompting the Commission's call for government-imposed legal obligations.

Key takeaway

For election officials and policymakers evaluating AI's role in democratic processes, this report underscores an urgent need for regulatory action. You should consider implementing legal obligations on AI platforms to prevent the spread of election misinformation, mirroring the UK Electoral Commission's push. Proactive regulation is crucial to safeguard electoral integrity and ensure voters receive accurate, verifiable information, especially given high voter reliance on AI tools.

Key insights

AI chatbots exhibit significant error rates in providing election information, necessitating regulatory oversight.

Principles

• AI tools frequently generate election misinformation.
• Voter reliance on AI for election data is substantial.
• Official source attribution is often absent in AI responses.

Method

A Demos report tested five free AI tools by asking 75 distinct election questions to assess misinformation rates.

In practice

• Evaluate AI tools for election information accuracy.
• Identify specific misinformation types (e.g., fake candidates).

Topics

• AI Regulation
• Election Misinformation
• Chatbots
• Voter Information
• UK Electoral Commission
• Demos Report

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Policy Maker, Legal Professional, AI Ethicist

Related on AIssential

• ChatGPT and other AI bots made huge errors before Scottish election, study finds — AI chatbots frequently generate election misinformation, necessitating urgent regulatory intervention to protect democratic processes.
• AI chatbots giving misleading voting advice in run-up to election - BBC — AI chatbots can provide inconsistent and factually incorrect political information, including voting recommendations and candidate details.
• Poll shows 71% of Americans think AI is progressing too quickly — The rapid pace of AI development is causing widespread public concern over job displacement, transparency, and regulation.
• An update on our election safeguards — AI models can be a positive force in democratic processes by providing accurate, impartial, and balanced election information.
• Artificial intelligence raises profound moral questions — for all of humanity to answer — AI's profound moral questions necessitate independent regulation and democratic public governance, not corporate self-regulation or "participation-washing."
• AI chatbots can effectively sway voters – in either direction — AI chatbots effectively sway political opinions by generating numerous factual claims, even if some are inaccurate.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.