Meta is training its AI using an entire nation’s data… with no opt-out

The question of how to train and improve AI tools is one that triggers fierce debate, and this is something that has come into sharp focus as It becomes clear just how Meta is teaching its own artificial intelligence.

The social media giant is -- perhaps unsurprisingly to many -- using data scrapped from Facebook and Instagram posts, but only in Australia. Why Australia? Unlike Europe where General Data Protection Regulation (GDPR) necessitated Meta to give users a way to opt out of having their data used in this way, Australia has not been afforded this same opportunity. What does this mean?

See also:

• Meta reveals more details of WhatsApp and Messenger's interoperability with third-party messaging services that's coming to Europe soon
• Microsoft builds on past Windows 11 taskbar improvements with enhanced app pinning
• Launch multiple apps at once with the new Workspaces module in Microsoft PowerToys v0.84.0

Let’s be clear from the very start, Australians have not been singled out by Meta as having data that is particularly useful or as people who are ripe for abuse. The company would much rather be left to do whatever it feels like doing with everyone’s data from all over the world. But, sadly for Meta, privacy rules mean users must be given the chance to opt-out of having their data used in this way, and -- predictably -- many people jump at this chance. It is the reason the company was forced to pause the rollout of its AI tools in Europe while it built an opt-out feature.

Similar rules do not exist in many other parts of the world, including Australia. So, with no good reason -- morals and ethics be damned -- to stop the wholesale harvest of users data in places where is it not explicitly banned from doing so, Meta is busy scooping up as much as it can while it still has the opportunity.

The state of play in Australia has been highlighted this week during a parliamentary committee exchange. As part of an inquiry into AI in Australia, Meta was asked why Australians had not been given the same opt-out choice as Europeans. Meta says that the option was introduced in Europe “in response to a very specific legal frame” -- i.e. it had no choice but to give people the choice.

Meta has tried to play things down by pointing out that it is only data that is public which is gathered, but this is the default setting that is in place for millions of users. The amount of data available to Meta is colossal, with scraping of public posts stretching as far back as 2007.

The chair of the inquiry, senator Tony Sheldon, put the following to Meta’s director of privacy policy, Melinda Claybaugh:

The truth of the matter is that, unless you consciously had set those posts to private, since 2007, Meta has just decided you will scrape all of the photos and all of the text from every public post on Instagram or Facebook that Australians have shared since 2007, unless there was a conscious decision to set them on private. But that’s actually the reality, isn’t it?

Her response? “Correct”.

Meta says that it currently has no plans to give an opt-out option to Australians. This means that not only are Facebook and Instagram users in Australia going to be exploited by Meta unless they know what steps to take, but there are implications for AIs themselves. Having training data skewed by different gathering principles around the world will unavoidably introduce bias and weighting, and that’s not beneficial for anyone.

Image credit: Cttpnetwork / Dreamstime.com