Meta wants to train its AI with its users’ data 😡

Cliquez ici pour lire en français

Meta, the parent company of Facebook and Instagram, is currently facing criticism over data protection. And for good reason: it has announced its intention to use its users’ posts and photos to train its artificial intelligence, without their explicit consent.

Automatic Notification and a complex Refusal Option 😡

Facebook and Instagram users should be aware that their posts and photos may soon be used to train Meta’s artificial intelligence (AI). From June 26, Meta’s generative AI « experiments » will be able to draw on all public content posted on Facebook and Instagram. A colossal amount of textual and visual data that will feed the group’s large language models (LLMs).

While users have the option of objecting to this data collection, the process is far from simple. They will have to fill in an online form explaining their objections, then confirm on their mailbox. However, Meta has made it clear that it reserves the right to use certain information despite users’ objections. An obstacle course that has drawn criticism from privacy advocates.

This complexity has been criticized by data protection experts such as Simon McGarr, Director of Data Compliance Europe. He points out that the process potentially violates the provisions of the European Union’s RGPD.

“Currently they’re throwing up a consent form. But the consent form doesn’t say, ‘yes I approve’. It’s fill in this form if you don’t approve. So it’s an opt out form,” said Simon McGarr, a solicitor and director of Data Compliance Europe, a company specialising in RGPD.

Legal and ethical issues ⚖️

According to the GDPR, personal data must be processed on a lawful and appropriate basis. Simon McGarr notes that sensitive data, such as that relating to sexual orientation or medical conditions, cannot be processed on the basis of legitimate interest, as Meta claims. Article 9 of the GDPR explicitly prohibits the processing of such data without explicit consent.

In addition, the Data Protection Commissioner (DPC) expressed her concerns, leading to a delay in the launch of this new policy. Meta has responded by introducing additional transparency measures and a dedicated objection mechanism. Nevertheless, questions remain as to the legality of this data processing.

Implications and Reactions of Regulators 👨🏾‍⚖️

Regulators, particularly in Europe, are keeping a close eye on the situation. The CPD has shared its concerns with its European counterparts and continues to interact with Meta to ensure compliance with data protection standards.

This situation is not unique to Meta. Other platforms, such as Reddit, have also failed to ask users for consent to use their data for AI development. This has already led to notable mistakes, such as Google’s AI Overview. To find out more, click here.

However, Meta’s new practice is fuelling fears about the potential misuse of personal data by major technology companies to feed their ever-greater artificial intelligence systems. This is a major issue that could be subject to stricter regulation in the near future.

Do you use Facebook or Instagram? If so, what do you think of this measure? Tell us in the comments.