It might be tempting to give social media giant Meta credit for walking back an AI feature that caused so much anger. But upon closer inspection, the company should have known better.
Shortly after introducing a new AI model, known as Muse Image, into its Meta AI and Instagram apps, users found out they were automatically opted into a feature that would allow others to generate AI content based on the photos they had uploaded. Suddenly the internet was flooded with content explaining how users needed to change their settings within the app to protect their photos – and therefore their identities – from being used for content that was beyond their control.
Meta, which owns Facebook, Instagram, WhatsApp and Threads, already has a lengthy history of proactively making decisions that have come under intense scrutiny from users. Those complaints have come in the form of not being fully transparent in how customer data is used, tracked and sold.
Other worries, focusing more on the endless scrolling interfaces of the various platforms, have become litigious, and in some cases caused the company to suffer embarrassing losses in court. Ultimately, the company’s recent controversial decision related to Muse Image led to it remove the scrutinised AI feature several days after rolling it out.
A statement reported by several media outlets from Meta sought to sooth the anger.
“Earlier this week, we announced that one way for people to generate images in Meta AI is by @-mentioning public Instagram accounts that they want to reference,” the company’s statement read in part. “Our intent was to provide a useful creative tool and to give people control over whether their public content could be referenced in this way. We’ve heard the feedback that this feature missed the mark, so it’s no longer available.”
“Missed the mark” might just be the understatement of the year.
For more than a year now, polls have shown an increased uneasiness from the public, especially in the US, when it comes to AI. Given the increasingly overt concern being expressed, it’s almost impossible for anyone at Meta, especially executives and product managers, to claim that they’re surprised by the reaction of those who didn’t want to automatically be opted into Muse Image’s AI feature.
All that aside, there’s a greater contradiction that’s become increasingly apparent from the California-based social media company trying to turn to AI.
Last year, in the early stages of trying to become a major player in the AI sector, Meta announced that it would be hiring Alexandr Wang to be the company’s chief AI officer. To the lay person, Mr Wang’s move to Meta didn’t have much impact at the time, but for those in technology circles, it was a clap of thunder.
That’s because as part of the deal, Meta would be taking a 49 per cent stake in Scale AI, the company Mr Wang co-founded. The AI data-labelling firm had quietly become one of the most important companies that most hadn’t heard of. Its origin story is the stuff of Silicon Valley legends. Mr Wang dropped out of the Massachusetts Institute of Technology to start Scale AI.
At an event hosted by the Centre for Strategic and International Studies last year, Mr Wang said he was ahead of the curve in seeing that AI would be one of the most important technology waves of our lifetime. It was through a sense of altruism, he told CSIS, that he wanted to devote himself to making sure the US – and not China – would be the world’s AI leader.
Mr Wang said that his upbringing in Los Alamos, New Mexico, made famous by Robert Oppenheimer’s development of the atomic bomb, informed his decision that the US, being on the right side of history, should win the AI race instead of China. “I went on a trip to China, one of our investors organised this trip, and we visited the Chinese AI companies at the time,” Mr Wang recalled. “100 per cent of them were focused on facial recognition and surveillance.”
It was at that moment, he said, that he knew AI would have national security implications, and the US would need to dictate how the AI sector would develop to ensure that it would be used properly. “We had, frankly, a moral imperative to support it,” Mr Wang concluded.
The entrepreneur would repeat similar versions of this outlook at several events and to a number of media outlets. Yet fast forward to earlier this month, and it hits with incredible hypocrisy, especially given how Meta decided to introduce Muse Image to the world.
All of a sudden, millions of Instagram users became aware that at any given time, by default, their images – and for that matter, their identity – could be used by someone else in another piece of AI content, without their consent. That is akin to the facial recognition and surveillance fears that supposedly informed Mr Wang’s opinion that the US, and not anybody else, should be the country to lead the world in AI research – therefore setting a peaceful and non-nefarious AI standard that would be embraced by just about everyone.
Sure, Meta disabled the controversial feature wherein users were automatically opted in, but the company has enough of a track record and ample knowledge of overall fears related to AI to know better than to implement it in the first place. They should have known better, but seemed to throw caution to the wind at a time when confidence in the technology sector leaves something to be desired.
Meanwhile, just one day after Muse Image was introduced, Mr Wang shared a post on X highlighting that Meta’s stock price rose sharply because of the company’s latest AI models. “New benchmark just dropped,” he wrote, not referring to any of the controversy, but instead implying that the stock price justified all of Meta’s AI efforts.
So much for moral imperatives.
