Social media giant, Meta Platforms has admitted it scrapped every Australian adult’s public images, texts, and other data posted on either Facebook or Instagram to train its AI models, the company revealed during an inquiry.
Unlike in the EU where laws related to AI are clear, Australia still holds on to privacy laws last reviewed in 2020 and deemed “outdated.” This has exposed Australians to manipulation by tech firms like Meta that scrap data from people’s accounts without their consent.
Meta used posts from as far back as nearly 20 years ago
During the inquiry, Meta’s global privacy director Melinda Claybaugh initially rejected the claims that the company scrapped data off people’s posts on Facebook and Instagram to train its AI models.
Upon being pressed by the lawmakers, Claybaugh eventually admitted the practice. Initially, Labor Senator Tony Sheldon asked if the social media giant had used Australian posts from as far back as 2007 to train its AI models, which Claybaugh denied.
However, Greens senator David Shoebridge chipped in challenging her. He asked:
“The truth of the matter is that unless you have consciously set those posts to private since 2007, Meta has just decided that you will scrape all of the photos and all of the texts from every public post on Instagram or Facebook since 2007, unless there was a conscious decision to set them on private. That’s the reality, isn’t it?”
Claybaugh admitted to the allegation responding with a simple “correct.” However, according to Claybaugh, accounts of users below the age of 18 were spared although their images would be scrapped if posted on an adult’s account. Meta is not new to controversies involving breaches of users’ privacy.
In June, a group of journalists, photographers, and curators led by Pulitzer award-winner Daniel Etter signed an open letter to stop the social media giant from training its AI models on real images of war, conflict, and crisis.
This followed revelations by Meta that it has an advantage in the generative AI space because of all the “public” photos.
Meta doesn’t give Australians an opt out option
Unlike in the EU, Meta does not give an opt out option for users in Australia. Claybaugh said the company does so in the EU due to its regulations that is strict on such. In Australia, the laws do not mandate the company to offer an opt out option.
“In Europe there is an ongoing legal question around what is the interpretation of existing privacy law with respect to AI training,” she said.
Due to the strict laws in the EU, Claybaugh revealed the company had halted introducing some of its AI products as the environment is also marred in uncertainty.
However, while Australians do not have the opt out option, Claybaugh revealed that, they could set their data to private. The development comes head on the heels of federal government’s threats, vowing to introduce bans on social media for children due to the harm the platforms were causing on minors.
Senator Shoebridge however insinuated there is reluctance on the part of government in coming up with laws to protect Australians and their privacy.
“The government’s failure to act on privacy means companies like Meta are continuing to monetize and exploit pictures and videos of children on Facebook.”
Shoebridge.
“There’s a reason that people’s privacy is protected in Europe and not in Australia, it’s because European lawmakers made tough privacy laws. Meta made it clear today that if Australia had these same laws, Australians’ data would also have been protected,” he told the ABC.
However, the government is expected to announce amends to its Privacy Act, following a 2020 review that showed the laws were now outdated.
Source: https://www.cryptopolitan.com/meta-admits-to-using-users-data-to-train-ai/