The European Data Protection Board (EDPB) issued, on Tuesday 17 December, an Opinion on the processing of the personal data of users of networks or platforms for the purpose of training artificial intelligence models.
This issue has come to the fore on several occasions in recent months, most notably in the cases of X, Meta, Apple and Google, all four of which have been asked by the EU to give up using the personal data of their European users to train their AI models (see EUROPE 13481/11).
In its Opinion, the EDPB “considers that AI models trained with personal data cannot, in all cases, be considered anonymous” and that each model claiming this should be examined in the light of its specific case.
However, the Committee is not opposed to the use of such data for this purpose and believes that the argument of legitimate interest can be a legal basis for justification.
It considers that where the interests, rights and freedoms of the data subject “seem to override the legitimate interest(s) being pursued by the controller or by a third party”, the controller may consider introducing “mitigating” measures to limit the impact of the processing on these data subjects.
However, these measures should not be confused with mandatory measures to ensure compliance with data protection legislation.
To see the opinion: https://aeur.eu/f/ew5 (Original version in French by Isalia Stieffatre)