Meta training AI with Facebook Data

 

Meta is facing multiple complaints in Europe over plans to train AI on user data. It seems users have the right to opt out but without opting out, users' data will be used to train the Meta AI. 

User's posts will include images and comments. Private messages used in Messenger should be excluded at this stage. Legally, it is a tough challenge to argue what information can be used to train or not Meta AI, considering all the data sits on Meta's databases and they are not pulled from other resources. 

Open AI Chat GPT scraped the web much before Meta and it was in fact part of legal actions, specifically, for the content scraped on copyrighted materials such as books. However, when it comes to Facebook, it can be very challenging to define its limits, and like for ChatGPT, you would probably notice a copyright or privacy infringement when its too late.

So far spokesman from Meta have replied to accusation by pointing at other techs and how they got away without having to deal with privacy laws. Meta also stated that they will comply with privacy laws but backed up their statement saying their approach won't be different to what other tech companies are doing.