Meta detailed its plans to train its models on UK adult users' public content

Meta recently shared a blog post in which it announced it will reinstate the training of its AI models on public posts, pictures, captions, and comments by UK-based adult users on Facebook and Instagram starting this week. The post reinforces the narrative in which Meta frames itself as taking essential and irreplaceable actions to ensure that its AI products accurately represent the culture, history, and idioms of every region in which they are available, and welcomes the UK's Information Commissioner’s Office (ICO) support of Meta's decision to justify its data collection methods as under the basis of ‘Legitimate Interests’.

This is not the first time Meta has attempted the 'legitimate interests' justification, but it is the first time that it has worked in its favor. Previously, the Court of Justice of the European Union rejected Meta's claim that processing data for targeted advertisements fell under a 'legitimate interest' to provide the best service possible. Moreover, the non-profit noyb– European Center for Digital Rights has filed complaints in 11 European countries noting, among other things, that the company is following the same strategy to justify overriding the privacy and rights of users in an even broader and more aggressive manner. Shortly after noyb filed its complaints, Norway's data protection agency published a blog post arguing that Meta's approach was unlikely legal. AI training on public posts is still paused in the European Union.

Meta did implement some noticeable changes in how it informs users about the data collection process and the ease with which users can opt-out. The company will send in-app notifications to eligible users, explaining what Meta is doing and the process they need to follow to opt their data out. Users who already have filled out the objection form will not receive these notifications, as Meta will continue to honor those requests and any others it receives.