Artificial intelligence (AI) assistants have become integral to our daily lives, assisting with tasks ranging from setting reminders to answering questions. However, a recent study has shed light on a concerning issue: these AI assistants are making frequent and widespread errors when reporting news.
The Study: A Deep Dive into AI’s Accuracy
A new study published on October 22, 2025, has uncovered alarming issues with AI assistants. The research, conducted by the European Broadcasting Union (EBU) in collaboration with the BBC, analyzed over 3,000 news-related queries answered by leading AI assistants like ChatGPT, Google’s Gemini, Microsoft Copilot, and Perplexity.
The study focused on how AI handles news reports and examined the accuracy, sourcing, and fact-checking abilities of these systems. The findings were shocking:
- 45% of responses contained at least one significant error.
- 81% had some form of issue, ranging from minor inaccuracies to major factual mistakes.
Common Errors Identified
The study highlighted several recurring problems:
- Sourcing Errors: Approximately 31% of responses had incorrect or missing citations, making it difficult to verify the information provided.
- Factual Inaccuracies: Around 20% of answers contained outdated or incorrect information. For instance, some AI assistants continued to refer to Pope Francis as the current pontiff months after his passing.
- Misinformation: Instances were found where AI assistants fabricated news links or provided misleading summaries of events.
Gemini’s Performance Under Scrutiny
Among the AI assistants evaluated, Google’s Gemini exhibited the highest rate of errors:
- 72% of its responses had significant sourcing issues.
This raised concerns about the reliability of information provided by Gemini, especially given Google’s prominence in the AI and search engine markets.
Implications for Public Trust
As AI assistants increasingly replace traditional search engines for news consumption, the implications are profound. The study’s authors warn that the proliferation of inaccurate information could erode public trust in digital platforms and hinder democratic engagement.
Jean Philip De Tender, Media Director at the EBU, emphasized the potential dangers: “When people don’t know what to trust, they end up trusting nothing at all, and that can deter democratic participation.”
The Rise of AI in News Consumption
The shift towards AI-driven news consumption is evident. According to the Reuters Institute’s Digital News Report 2025:
- 7% of all online news consumers and
- 15% of individuals under 25 now rely on AI assistants for news.
This trend underscores the urgency for AI developers to enhance the accuracy and reliability of their systems.
Calls for Accountability and Improvement
In response to the findings, the EBU and BBC have called for greater accountability from AI companies. They urge developers to:
- Enhance accuracy in news reporting.
- Provide clear sourcing for information.
- Distinguish clearly between fact and opinion to prevent misinformation.
The study also highlights the need for transparency in AI development and the importance of regular audits to ensure the integrity of information provided to users.
Conclusion
While AI assistants offer convenience and efficiency, this study serves as a stark reminder of their current limitations in delivering accurate news. As reliance on these technologies grows, it becomes imperative for developers to address these issues to maintain public trust and ensure the responsible dissemination of information.
For more details on the study, you can refer to the full report by the European Broadcasting Union and the BBC.
