Think twice before trusting your favorite AI chatbot with news. According to latest research , 45 percent of responses to questions about the news had at least one ‘significant’ issue. The shocking findings challenge everything we thought we knew about artificial intelligence misinformation.
The study published by the European Broadcasting Union (EBU) and the BBC assessed the accuracy of more than 2,700 responses given by OpenAI’s ChatGPT, Google’s Gemini, Microsoft’s Copilot, and Perplexity. This marks the largest investigation of its kind. You won’t believe what researchers discovered.
Alarming Results Shake Public Trust in AI Systems
AI models misrepresent news in ways that threaten democracy itself. The research exposed that 45% of analyzed AI responses contained significant issues, with 81% having some form of error, notably in sourcing and distinguishing between opinion and fact. These aren’t minor glitches. We’re talking about fundamental problems that could reshape how people understand current events.
The international research studied 3,000 responses to questions about the news from leading artificial intelligence assistants – software applications that use AI to understand natural language commands to complete tasks for a user. It assessed AI assistants in 14 languages for accuracy, sourcing and ability to distinguish opinion versus fact. The scope is unprecedented.
Meanwhile, EBU media director Jean Philip De Tender warned that “When people don’t know what to trust, they end up trusting nothing at all, and that can deter democratic participation”. This isn’t just about technology anymore.
AI Chatbot Accuracy Issues Plague All Major Platforms
AI chatbot accuracy suffers across every major platform. The headline figures are sobering: 45% of responses contained at least one significant issue, while 81% had some form of problem (including minor issues). No system emerged unscathed.
Google’s Gemini performed particularly poorly. Google’s Gemini registered a disproportionately high rate of sourcing problems in the EBU data: a reported ~72% of Gemini responses in the sample had significant sourcing issues, compared with under 25% for other assistants in the panel. Those numbers are staggering.
This was one of the largest evaluations of its kind to date, including 22 public service media organizations across 18 countries, and in 14 languages. The study’s results found that AI assistants routinely misrepresent news content no matter which language, territory, or AI platform is tested.
How AI Models Misrepresent News: The Technical Breakdown
The ways AI models misrepresent news follow disturbing patterns. Sourcing was the most common problem, with 31 percent of responses including information not supported by the cited source, or incorrect or unverifiable attribution, among other issues. A lack of accuracy was the next biggest contributor to faulty answers, affecting 20 percent of responses, followed by the absence of appropriate context, with 14 percent.
These systems fail in three critical ways:
- Source attribution errors: AI models cite non-existent sources or misattribute information
- Factual inaccuracies: Models present outdated or plainly false information as current
- Context stripping: Complex news stories lose nuance and become oversimplified
Misattribution or missing sources — assistants presenting authoritative‑sounding claims without traceable citations. Altered or fabricated quotes — paraphrasing or inventing attributions that change the original meaning. Context stripping and editorialisation — converting hedged, cautious reporting into assertive summaries that exaggerate or misrepresent.
Furthermore, artificial intelligence misinformation spreads through what experts call “confident prose syndrome.” The EBU/BBC international study is a wake‑up call for anyone who relies on conversational AI for news: these systems frequently misrepresent reporting in ways that matter, and the problem is concentrated in sourcing, context and the translation of hedged journalism into confident prose.
Real-World Examples of AI News Misrepresentation
Consider these troubling examples from the study. Google Gemini was found to misrepresent NHS vaping recommendations, a critical health topic where misinformation could have significant consequences. Lives could be at risk from such errors.
A particular concern highlighted in the study involves AI chatbots citing outdated or incorrect details about political leadership, as was the case with ChatGPT and Microsoft Copilot in mentioning political figures in the UK. Political misinformation undermines democratic processes.
Similarly, Google’s Gemini misrepresented recommendations from the NHS regarding vaping, while Perplexity AI provided inaccurate quotes from BBC’s Middle East coverage. Each error compounds the problem of trust.
The Growing Threat to Journalism and Democracy
These findings arrive at a critical moment for media consumption. Some 7% of all online news consumers and 15% of those aged under 25 use AI assistants to get their news, according to the Reuters Institute’s Digital News Report 2025. Young people rely increasingly on these flawed systems.
AI assistants have moved from novelty to default for many users: conversational answers are increasingly replacing the click‑through to primary reporting, and a growing minority of people now use AI for news. The Reuters Institute’s Digital News Report 2025 estimates that around 7% of online news consumers (and 15% of those under 25) rely on AI assistants for news — a nontrivial audience that makes the accuracy of assistant answers a public interest issue.
The implications extend far beyond individual misinformation. Such public reactions highlight a broader anxiety about the implications of AI errors on democratic processes and societal trust.
Public Response and Industry Accountability
People are rightfully concerned about these revelations. Social media platforms have seen an outpouring of concern regarding AI’s capability to distort facts and generate misleading narratives. Particularly, the public appears apprehensive about AI’s tendency to fabricate quotes or misrepresent authoritative information, as noted in the BBC study.
The industry must respond with concrete action. The EBU report urged AI companies to improve how their AI assistants respond to news-related queries and to be more accountable, citing the example of how news organizations themselves have “robust processes to identify, acknowledge and correct” errors. “It is important to make sure that the same accountability exists for AI assistants,” it said.
However, Reuters has made contact with the companies to seek their comment on the findings. Many companies have yet to provide substantive responses to address these critical issues.
Understanding the AI News Reliability Study Methodology
The AI news reliability study employed rigorous scientific methods. Twenty-two public media outlets, representing 18 countries and 14 languages, posed a common set of questions to the AI assistants between late May and early June for the study. Researchers left no stone unturned.
Journalists and subject experts from 22 public broadcasters in 18 countries reviewed outputs for a range of failures: factual inaccuracy, missing or incorrect sourcing, editorialisation, loss of context and failure to distinguish opinion from verified fact. Human expert review: journalists and subject experts reviewed outputs, making judgments based on editorial standards rather than an automated truth metric. That aligns the evaluation with how real audiences and editors judge news quality.
The study’s design sets new standards for evaluating AI systems. Multi-language and cross-platform: the study tested assistants across 14 languages and multiple countries, increasing generalisability beyond English-only evaluations. Multiple error categories: the reviewers recorded nuanced errors (attribution, editorialising, accuracy, context), enabling more targeted analysis than a binary “right/wrong” test.
Chatbot False Information Rate Reveals Systemic Problems
The chatbot false information rate reveals deeper issues than simple mistakes. Some publishers restrict crawlers or licensing, which means models must rely on second-hand citations or noisy web copies. If an assistant can’t access the canonical article, it may infer a source or reconstruct quotes incorrectly. That contributes to the sourcing errors flagged by the study. This is a systems problem — a mismatch between legal access, retrieval design and output summarisation.
The study’s cross-product comparison produced striking patterns rather than a single winner. Key takeaways: No assistant was free of problems. Even systems with strong retrieval capabilities produced contextual errors.
These aren’t bugs that can be easily fixed. They represent fundamental challenges in how AI models process and present information.
Solutions and Future Implications
Despite these challenges, solutions exist for responsible AI development. While AI assistants deliver clear benefits — speed, accessibility and new discovery patterns — their integration into news delivery demands higher engineering standards, editorial guardrails and transparent provenance.
The study provides a practical diagnostics framework and immediate policy levers that vendors, publishers and regulators can adopt to protect news integrity. Until those fixes are widely implemented, the best practice for Windows users and all news consumers is straightforward: use AI assistants for leads and discovery, not as a substitute for primary reporting; always check the sources.
Companies must invest in better training data and verification systems. Journalists need to reclaim their role as information gatekeepers. Users require media literacy education to navigate this complex landscape.
The path forward demands collaboration between technologists, journalists, and policymakers. AI models misrepresent news because we’ve allowed them to operate without sufficient oversight. That changes now.
Protecting Democracy Through Better AI Governance
The stakes couldn’t be higher for democratic societies. With AI assistants increasingly replacing traditional search engines for news, public trust could be undermined, the EBU said.
We need immediate action on multiple fronts:
- Transparency requirements for AI training data and methodology
- Accuracy standards that match traditional journalism ethics
- Source verification systems that validate information before presentation
- User education about AI limitations and bias
The research demonstrates that AI models misrepresent news events nearly half the time. This isn’t acceptable for systems that shape public understanding of current events.
The Road Ahead: Rebuilding Trust in AI Information Systems
Moving forward, the industry must confront these findings honestly. Looking ahead, the implications of AI inaccuracies in news reach far beyond the technology itself. We’re talking about the foundation of informed democratic participation.
In light of this, we have coordinated the European rollout of a BBC study into how AI assistants are distorting the answers to news queries with 24 Members agreeing to replicate the study in their own territories. International cooperation is essential.
The findings should serve as a wake-up call rather than a reason for despair. AI technology offers tremendous potential for enhancing journalism and public information access. However, realizing that potential requires acknowledging current limitations and working systematically to address them.
Companies developing AI models misrepresent news must prioritize accuracy over speed. Publishers need to maintain editorial standards even when partnering with AI platforms. Users deserve transparent information about system capabilities and limitations.
This groundbreaking study marks a turning point in our understanding of artificial intelligence misinformation. The question isn’t whether AI will play a role in news distribution – it already does. The question is whether we’ll build systems worthy of public trust or continue accepting mediocrity that threatens democratic discourse.
The choice is ours, but time is running short.
FAQs:
Q1: How often do AI models misrepresent news according to the study?
A: The EBU-BBC study found that AI models misrepresent news events 45% of the time, with 81% of responses containing some form of error.
Q2: Which AI assistants were tested in the news accuracy study?
A: Researchers tested ChatGPT, Google Gemini, Microsoft Copilot, and Perplexity across 14 languages and 18 countries.
Q3: What are the main problems with AI chatbot accuracy in news reporting?
A: The primary issues include sourcing errors (31%), factual inaccuracies (20%), and lack of appropriate context (14%).
Q4: Which AI assistant performed worst in the study?
A: Google’s Gemini showed the highest rate of problems, with 72% of responses having significant sourcing issues compared to under 25% for other assistants.
Q5: How many people use AI assistants for news consumption?
A: According to the Reuters Institute, 7% of all online news consumers and 15% of those under 25 use AI assistants for news.
Q6: What solutions does the study recommend for AI news reliability issues?
A: The study recommends higher engineering standards, editorial guardrails, transparent sourcing, and using AI for discovery rather than as a substitute for primary reporting.
Q7: Why is AI news misrepresentation a threat to democracy?
A: When people don’t know what information to trust, they may end up trusting nothing at all, which can deter democratic participation and undermine informed decision-making.
