Meta Wants To Use Facebook And Instagram Posts From As Far Back As 2007 To Train Its AI


AI is on the rise, and tech companies are doing anything they can to compete with the lucrative market.

Meta, the tech titan that is also the largest social media company the world has ever seen, is picking up its pace in the generative AI race. Since OpenAI introduced ChatGPT, Meta has been one of the many companies that entered the race, trying to be the best out of the best.

But in order to stay competitive, Meta has to get its hands on a lot of data to train its AI.

While most others resort to the internet itself, Meta, being the social giant, is able to tap into whatever is in its servers, to train its AI.

To use its platform, Meta asks people their consent, and to agree to the terms of service.

The process is easy, and most of the time, people don't really much care, and never actually read it.

And here, Meta has updated its policy, to say in a website post that it's starting to use information that include users' social activities on Facebook and Instagram, to train its AI tools.

According to the change, data from as far back as 2007 shall be used, and they include everything from social media posts, photos, captions, their interactions with pages, and messages users made with Meta's AI chatbot.

And thanks to Meta's unrivaled reach, the company is also gathering data from third parties which use its codes.

What this means, Meta also wants to collect information about people who aren't on Facebook or Instagram.

And data collected for AI training could be shared with third parties.

The only exception from AI training is made for private messages sent between "friends and family," which will not be processed, Meta's blog said.

In other words, starting June 26, when that policy comes into effect, Meta would make pretty much all users' interactions to train and improve its AI.

Meta notif

For certain, the move quickly raised privacy concerns and complaints, which argues that Meta’s use of personal data for AI training constitutes misuse.

For example, the European Center for Digital Rights, known as Noyb, has filed complaints (PDF) in 11 European countries to halt Meta's plan to start training its AI on European Union-based Facebook and Instagram users' personal posts and pictures.

The controversy centers around Meta’s intention to use both public and non-public user data for an undefined range of current and future AI technologies.

"Unlike the already problematic situation of companies using certain (public) data to train a specific AI system (e.g. a chatbot), Meta's new privacy policy basically says that the company wants to take all public and non-public user data that it has collected since 2007 and use it for any undefined type of current and future 'artificial intelligence technology,'" Noyb alleged in a press release.

And "once their data [is] in the system," Noyb said, "users seem to have no option of ever having it removed."

Making things worse, according to Noyb, is Meta's overly complicated opt-out process.

Part of Meta's "deal" is to provide ways for Facebook and Instagram users to opt out of AI training data sets. But Noyb criticized Meta for not simply providing a one-click option, and instead requiring users to log in to access a public form and provide a reason for opting out.

According to Noyb, Meta's "excessive use" of dark patterns starts with the email sent notifying users of the policy change.

Any users who actually opened the email and clicked the link to opt out were redirected to a login page, after which they were redirected to their newsfeed rather than the opt-out form.

"Data subjects therefore had to go back to the email and click the link a second time (while now being logged in) to even reach the form," Noyb's complaint said.

Meta does provide a opt out form, but to find that form, Meta users would have had to scroll to the bottom of Meta privacy policy page for generative AI, Noyb found.

At this time, only users in the European Union and the U.S. state of Illinois can opt out, because they have AI protection laws like the General Data Protection Regulation (GDPR) in place.

"We keep training data for as long as we need it on a case-by-case basis to ensure an AI model is operating appropriately, safely, and efficiently," one Meta AI privacy page said. "We also may keep it to protect our or other’s interests or comply with legal obligations."

This AI initiative is part of Meta's effort to appeal to younger users.