Meta halts plan to train AI using European users’ data, bows to regulatory pressure


Meta has confirmed that it will halt plans to train its AI systems using data from its users in the EU and UK

The move comes after opposition from the Irish Data Protection Commission (DPC), Meta’s lead regulator in the EU, acting on behalf of a number of data protection authorities across the bloc. The UK’s Information Commissioner’s Office (ICO) also requested that Meta put its plans on hold until it addressed its concerns.

“The DPC welcomes Meta’s decision to halt its plans to train its large language models using public content shared by adults in the EU/EEA on Facebook and Instagram,” the DPC said in a statement on Friday. “This decision was taken following intensive dialogue between the DPC and Meta. The DPC, together with its fellow EU data protection authorities, will continue to engage with Meta on this issue.”

While Meta is already using user-generated content to train its AI in markets such as the US, Europe’s stringent GDPR regulations have created hurdles for Meta – and other companies – that want to improve their AI systems, including large language models, with user-generated training content.

However, Meta last month began notifying users about a coming change to its privacy policy, which it said would give it the right to train its AI using public content on Facebook and Instagram, including the content of comments, interactions with companies, status updates, photos and captions attached to them. The company argued that it needed to do so to reflect the “diverse languages, geography and cultural contexts of people in Europe”.

These changes were supposed to come into effect on June 26 – 12 days from now. But these plans prompted the nonprofit privacy activist organization NOYB (“None of Your Business”) to file 11 complaints with constituent EU countries, arguing that Meta is violating various aspects of the GDPR. One of them relates to the issue of opt-in versus opt-out, face to face Where personal data is processed, users must first be asked for their consent, rather than being denied it.

Meta, for its part, relied on a GDPR provision called “legitimate interest” to claim that its actions were in line with regulations. This is not the first time Meta has used this legal basis in defense, having previously done so to justify the processing of European users for targeted advertising.

It always seemed likely that regulators would at least put a halt to implementation of Meta’s planned changes, especially given how difficult the company made it for users to “opt out” of having their data used. The company said it sent more than 2 billion notifications to users to inform them of the upcoming changes, but unlike other important public messages that are pasted to the top of users’ feeds, like prompts to go out and vote, these notifications appear alongside users’ standard notifications: friends’ birthdays, photo tag alerts, group announcements, and more. So if someone doesn’t regularly check their notifications, it’s all too easy to miss.

And those who saw the notification did not automatically know there was a way to object or opt-out, as it simply invited users to click to learn how Meta would use their information. There was nothing to suggest there was a choice here.

Meta: AI Notification
Image Credit: Meta

Furthermore, technically users were not able to “opt out” of having their data used. Instead, they had to fill out an objection form where they had to submit their reasoning as to why they did not want their data to be processed – it was entirely at Meta’s discretion whether to accept this request or not, although the company said it would honor every request.

Facebook "Objection" Form
Facebook “Objection” Form
Image Credit: meta / screenshots

Even though the objection form was linked to the notification itself, it was still very difficult for anyone who was searching for the objection form in their account settings.

On the Facebook website, they first need to profile photo top-right; hit Settings and privacy; Tap Privacy Center; Scroll down and click on Generative AI in Meta section; scroll down again and see a bunch of links to a section titled more resourcesThe first link under this section is called “How Facebook uses information for meta-generative AI models,” and they had to read about 1,100 words before reaching a separate link to the company’s “right to object” form. It was the same story in the Facebook mobile app.

link to "Right to object" Form
Link to “Right to Objection” form
Image Credit: meta / screenshots

Earlier this week, when asked why the process requires users to register an objection rather than opt-in, Meta’s policy communications manager Matt Pollard pointed TechCrunch to its existing blog post, which states: “We believe this legal basis (“legitimate interest”) is the most appropriate balance for processing public data at the scale needed to train AI models, while respecting people’s rights.”

This meant that making it opt-in would not generate enough “scale” for people to give up their data. So the best way to do this was to issue a single notification among users’ other notifications; hide the objection form behind half a dozen clicks for those who wanted to “opt-out” independently; and then ask them to justify their objection, rather than giving them a straight opt-out.

In an updated blog post on Friday, Stefano Fratta, Meta’s global engagement director for privacy policy, said they were “disappointed” by the request they received from the DPC.

“This is a step backwards for European innovation, competition in AI development, and a further delay in delivering the benefits of AI to people in Europe,” Fratta wrote. “We are confident that our approach complies with European laws and regulations. AI training is not unique to our services, and we are more transparent than many of our industry counterparts.”

The AI ​​arms race

None of this is new, and the meta AI arms race has shone a huge light on the vast troves of data that Big Tech has on all of us.

Earlier this year, Reddit revealed it had signed contracts worth over $200 million to license its data to companies like ChatGPT-maker OpenAI and Google over the coming years. And the latter of these companies is already facing a hefty fine for relying on copyrighted news content to train its generative AI models.

But these efforts also highlight the lengths companies will go to to leverage this data within the boundaries of existing law; “opting in” is rarely on the agenda, and the process to opt out is often unnecessarily difficult. Just last month, someone spotted some questionable wording in the existing Slack privacy policy that suggested it would be able to leverage user data to train its AI systems, with users able to opt out simply by emailing the company.

And last year, Google gave online publishers the option to exclude their websites from the training of its models, allowing them to insert a chunk of code into their sites. OpenAI, for its part, is building a dedicated tool to allow content creators to exclude them from the training of its generative AI smarts; it should be ready by 2025.

While Meta’s efforts to train its AI on users’ public content in Europe are currently on hold, it will likely resurface in some form after consultation with the DPC and ICO – hopefully with a different user-permissioning process.

“To make the most of generative AI and the opportunities it brings, it is vital that the public can trust that their privacy rights will be respected from the start,” Stephen Almond, the ICO’s executive director for regulatory risk, said in a statement on Friday. “We will continue to monitor major developers of generative AI, including Meta, to review the safeguards they have put in place and ensure that the information rights of UK users are protected.”


Please enter your comment!
Please enter your name here

Share post:




More like this

Samsung Galaxy Z Flip 6 rumours: Release date, camera specs and more

The larger cover screen of the Galaxy Z Flip...

Airbnb’s Olympic campaign could lead to victory in Paris

Short-term rentals can act as a quick release valve...

Waabi’s GenAI promises to do more than power self-driving trucks

For the past two decades, Rakel Urtasun, founder and...

Lords of the Fallen sequel set for release in 2026, will be Epic Games Store exclusive on PC

A sequel to the rebooted Soulslike action-RPG Lords of...