Meta pauses plans to train AI using European users’ data after regulator pushback

Meta, the company behind Facebook and Instagram, has put its plans to train AI systems using European user data on hold. This follows pushback from data protection regulators in the EU and UK, citing concerns about user privacy under Europe's strict General Data Protection Regulation (GDPR).

Meta planned to train its AI models using publicly shared content from adult European users. Last month, the company began notifying users of an upcoming change to its privacy policy, one that would give it the right to use public content on Facebook and Instagram to train its AI. The content included user comments, interactions with companies, status updates, photos, and their associated captions.

The changes were set to take effect starting June 26. However the plan was cut short after a non-profit privacy advocacy group, NOYB (“none of your business”) filed complaints with various EU countries. While Meta argued this complied with regulations, the Irish Data Protection Commission (DPC) and the UK's Information Commissioner's Office (ICO) disagreed.

A key concern was Meta's reliance on an "opt-out" system, where users had to actively object to their data being used. Regulators argued that GDPR prioritizes "opt-in," requiring explicit user consent for data processing.

The regulators also criticized Meta's notification process for the privacy policy noting that it was buried within other notifications and lacked clear instructions on how to opt out.

In its defense Meta argued its approach is legal and highlights its transparency compared to other companies. The social media company said that it complied with the "legitimate interests" provision of GDPR that allows for using public data for AI training.

Meta, which is already using user-generated data to train its AI in the US and other markets, views the decision as a setback for European AI development. However, regulators emphasize the need to balance innovation with user privacy rights.

While Meta's plans are paused, consultations with regulators will likely lead to a revised approach with a more user-friendly opt-in system.

However, this case is not anything new. Ever since OpenAI started the AI revolution in 2020, one of the biggest contention points has been the data that big tech companies hold over us and the legal/moral standing in using this data to train AI systems.

Just recently, Reddit revealed that it has been offered up to $200 million by OpenAI and Google to license its content to them. OpenAI is already being sued for using copyrighted news content to train its system.