EU report finds 45% of AI answers have at least one significant issue in summarizing news stories in study that may as well be a full-throated endorsement of the Leviathan News aggregation model

A new report from the European Broadcasting Union and the BBC finds that AI assistants still struggle to reliably summarize news, with 45% of evaluated answers containing at least one significant issue and 81% showing some form of problem across accuracy, sourcing, or context. The study tested popular AI assistants on a set of news-related queries and found recurring failures such as hallucinated details, missing attribution, outdated information, and weak distinction between fact and interpretation, reinforcing concerns that these tools can mislead users when used as news gateways.

✨ AI-generated background, compiled from web sources — not editorial content.

More coverage

• Axel Voss' Good Intention – The European Parliament adopts report ... (legalblogs.wolterskluwer.com)
• New WHO/Europe report provides first-ever snapshot of AI in health ... (who.int)
• High-level summary of the AI Act | EU Artificial Intelligence Act (artificialintelligenceact.eu)
• [PDF] The Impact of Google AI Summaries and Google AI Overviews on ... (europarl.europa.eu)
• EU investigates Google over AI-generated summaries in search ... (news.ycombinator.com)

Explore the topic

• Leviathan
• EU
• AI