Italian Antitrust Investigates DeepSeek for Inadequate Hallucination Warnings
Italy's antitrust regulator opened an investigation into DeepSeek on 16 June 2025, alleging the AI chatbot failed to adequately warn users that its outputs could contain hallucinations—inaccurate, misleading or fabricated information. The Authority found only a generic English-language disclaimer visible to users and argued this could lead consumers to incorrectly trust DeepSeek's accuracy.
What happened
The Italian Antitrust Authority (Autorità Garante della Concorrenza e del Mercato) opened an investigation into Hangzhou DeepSeek Artificial Intelligence and Beijing DeepSeek Artificial Intelligence on 16 June 2025 for unfair commercial practices under the Consumer Code.
The Authority found that DeepSeek failed to inform users in a sufficiently clear, immediate and intelligible manner that AI model outputs could contain hallucinations—situations in which the AI generates inaccurate, misleading or fabricated information in response to user input.
DeepSeek displays only a brief generic English-language disclaimer reading 'AI-generated, for reference only' in dialogue windows, which the Authority deemed excessively generic and inadequate to inform consumers. No other warnings or disclaimers about hallucinations appear in the interface.
Reported impact
The lack of clear and accessible information could lead users to develop an incorrect belief that they can rely fully on the accuracy and correctness of DeepSeek's results.
Index analysis
DeepSeek was given 30 days to provide additional information, including the number of Italian users from launch until 1 March 2025 and a clear description of its practices.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.