AI News Assistants Struggle with Accuracy and Trust, New Global Study Reveals

Widespread Issues in AI-Generated News Content

A comprehensive international study has uncovered alarming deficiencies in how artificial intelligence assistants handle news content. The research, coordinated by the European Broadcasting Union and led by the BBC, examined more than 3,000 responses from major AI platforms including ChatGPT, Copilot, Gemini, and Perplexity. The findings reveal that nearly half of all AI-generated news responses contain significant errors or omissions that could mislead users seeking reliable information.

Widespread Issues in AI-Generated News Content
Concerning Statistics on AI Performance
Platform Performance Variations
Growing User Reliance on AI for News
Systemic Challenges and Industry Response
Moving Toward Solutions

Concerning Statistics on AI Performance

The evaluation assessed AI responses against critical journalistic standards including accuracy, proper sourcing, distinction between fact and opinion, and contextual completeness. The results demonstrated that 45% of all AI answers had at least one major issue, raising serious questions about the reliability of these tools for news consumption., according to technology trends

Breaking down the specific problems identified:, according to related news

Over 30% exhibited serious sourcing problems including missing, misleading, or incorrect attributions
20% contained major accuracy issues, including hallucinated details and outdated information
14% failed to provide sufficient context for proper understanding

Platform Performance Variations

Not all AI assistants performed equally in the assessment. Gemini emerged as the worst performer, with significant issues detected in 76% of its responses—more than double the rate of other platforms. The primary weakness identified was poor sourcing performance, particularly misattributing claims, which becomes especially problematic when those claims are factually incorrect., according to related news

The research also noted, earlier coverage, that AI assistants rarely refuse to answer questions, even when they cannot provide high-quality responses. Out of 3,113 questions posed during the study, only 17 (0.5%) received refusals—fewer than the 3% refusal rate observed in a previous BBC survey conducted in February.

Growing User Reliance on AI for News

These findings arrive at a critical moment when usage of AI for news consumption is rapidly increasing. According to the Reuters Institute’s Digital News Report 2025, 7% of online news consumers now use AI assistants to access news, with this figure rising to 15% among users under 25 years old.

Perhaps more concerning is the trust levels revealed in separate research. A BBC report found that approximately one-third of UK adults completely trust AI to produce accurate information summaries, with this figure approaching half among adults under 35.

Systemic Challenges and Industry Response

Jean Philip De Tender, EBU media director and deputy director general, emphasized that these issues represent systemic problems rather than isolated incidents. “This research conclusively shows that these failings are not isolated incidents. They are systemic, cross-border, and multilingual, and we believe this endangers public trust,” he stated, adding that “When people don’t know what to trust, they end up trusting nothing at all, and that can deter democratic participation.”

Peter Archer, BBC programme director for generative AI, expressed both optimism and concern: “We’re excited about AI and how it can help us bring even more value to audiences. But people must be able to trust what they read, watch and see. Despite some improvements, it’s clear that there are still significant issues with these assistants.”

Moving Toward Solutions

In response to these findings, the research team has developed a News Integrity in AI Assistants Toolkit designed to help address the identified problems. Meanwhile, the EBU and its member organizations are advocating for stricter enforcement of existing regulations concerning information integrity, digital services, and media pluralism at both EU and national levels.

The researchers stress that ongoing independent monitoring of AI assistants is essential given the rapid pace of AI development. They note that when users encounter errors in AI-generated news summaries, they often blame both news providers and AI developers—even when mistakes originate from the AI systems themselves.

As AI continues to transform how people access information, these findings highlight the urgent need for improved accuracy, transparency, and accountability in AI-generated news content to preserve public trust and informed democratic participation.