PARIS: Artificial intelligence assistants including ChatGPT provided incorrect information approximately half the time when answering news-related questions according to a major European public broadcasters study.
The research identified numerous error types including confusion between genuine news and parody content alongside date inaccuracies and completely fabricated events.
European Broadcasting Union investigators examined four popular AI assistants specifically OpenAI’s ChatGPT, Microsoft’s Copilot, Google’s Gemini, and Perplexity.
Across all platforms 45% of AI-generated responses contained at least one significant problem regardless of language or geographical origin.
One-fifth of all answers demonstrated major accuracy problems featuring hallucinated details and outdated information.
Google’s Gemini performed most poorly with significant issues affecting 76% of its responses more than double the error rate of other assistants primarily due to inadequate source verification.
Twenty-two public media organisations from eighteen mostly European countries submitted identical news questions to the AI systems between late May and early June.
Outdated information represented one of the most frequent problems identified within the 3000 collected responses.
When multiple European broadcasters asked “Who is the Pope?” ChatGPT, Copilot and Gemini all incorrectly answered “Francis” despite his death and replacement by Leo XIV at the time.
French radio station Radio France received Gemini’s bizarre response about Elon Musk having “an erection in his right arm” during a Nazi salute question after the system misinterpreted satirical comedy content.
European Broadcasting Union deputy director general Jean Philip De Tender and BBC head of AI Pete Archer jointly concluded that AI assistants remain unreliable for news consumption.
Young people increasingly turn to AI assistants for information despite these documented accuracy deficiencies.
A June Reuters Institute global report indicated 15% of people under twenty-five use AI tools weekly for news summaries. – AFP
Hence then, the article about eu study finds ai assistants unreliable for news with 45 error rate was published today ( ) and is available on Daily Sun ( Middle East ) The editorial team at PressBee has edited and verified it, and it may have been modified, fully republished, or quoted. You can read and follow the updates of this news or article from its original source.
Read More Details
Finally We wish PressBee provided you with enough information of ( EU study finds AI assistants unreliable for news with 45% error rate )
Also on site :
- Driver dies in fiery freeway crash on Christmas Eve near Sacramento
- Trump’s Christmas Eve calls with children asking about Santa’s whereabouts are steeped in partisan politics
- JCPenney Has a 'Stunning' White Sapphire Jewelry Set on Sale for Just $16, and It's Perfect for Every Occasion