Mistral AI—Mediapart investigation found Mistral AI trained models on pirated books via Library Genesis
A Mediapart investigation reported that Mistral AI trained its models on books obtained from Library Genesis (LibGen), a pirated-book database, citing internal emails from co-founder Guillaume Lample describing use of the LibGen corpus. French publisher Nouveau Monde Editions said it found more than 200 of its titles - about 20% of its catalog - in the LibGen data, and major French publishers including Gallimard, Seuil, PUF, Les Belles Lettres, Armand Colin and La Decouverte said their works were 'massively pillaged.' Mistral denied any agreement with Nouveau Monde and said its models are trained on 'public sources, licensed content, and internally generated data,' but has not published the training-data summary required under the EU AI Act since August 2025.
Scoring Impact
| Topic | Direction | Relevance | Contribution |
|---|---|---|---|
| AI Oversight | -against | secondary | -0.50 |
| Intellectual Property Ethics | -against | primary | -1.00 |
| Overall incident score = | -0.643 | ||
Score = avg(topic contributions) × significance (high ×1.5) × confidence (0.57)
Evidence (1 signal)
Mediapart: internal emails show Mistral co-founder used LibGen pirated-book corpus; publisher found 200+ titles
Mediapart's investigation cited internal emails from Mistral co-founder Guillaume Lample referencing use of the Library Genesis pirated-book database, and Nouveau Monde Editions founder Yannick Dehee said he found over 200 of his titles (about 20% of his catalog) in the LibGen data. Mistral denied any agreement with the publisher and said its training data is public, licensed, or internally generated.