Photo Credit: wikipedia
Most people post content to Facebook without even a second thought about the role that artificial intelligence (AI) or machine learning plays in their overall social media experience. Yet, it is AI-powered algorithms that determine what you see in your newsfeeds, how your images and posts are sorted and tagged, and what type of content is flagged as relevant or engaging for other users. Facial and image recognition is now a very important part of how Facebook analyzes content online – and often with unintended consequences.
Details of the Facebook image outage
For one day this summer, Facebook users had a brief insight into exactly what these AI-powered algorithms were thinking. That’s because Facebook suffered an “image outage” in which images failed to load entirely. Instead, users simply saw a few cryptic tags that the AI-powered algorithms had figured out from any posted content. Instead of a beautiful summer rose in a handcrafted vase, for example, all the AI-powered algorithm saw was “table, plant, flower, indoor.”
As might be expected, this unique insight into what the machines were thinking caused a mix of anger and resentment on other social media platforms. Some were suitably impressed by Facebook’s image recognition prowess, while others were horrified at what the AI algorithms thought about them (wait, does my computer really think I’m fat?)
Takeaways from the Facebook image outage
There are two big takeaways from all this, of course. One is that “social” networks are now remarkably “anti-social.” Cynics might point out that machines are now in control, manipulating us in subtle ways. What if you woke up to find that your beautiful vacation photos were tagged as “boring” by the Facebook AI? Or, what if you found that the AI algorithm had somehow made a major mistake about your race, gender, or ethnic identity? You might just find yourself “triggered” and give up social media entirely.
The other big takeaway is that Facebook might be tying its future just a little bit too much to the world of artificial intelligence. It was cute and endearing when Facebook was using AI to recognize the photo of your Uncle Bob in a photo you uploaded to the social network, but it’s a little creepy when Facebook can now glean all sorts of details about your private life from a seemingly innocuous photo you post online.
It might not be quite the dire AI scenario that tech luminaries like Elon Musk and Bill Gates have described (in which AI-powered machines take over from humanity and enslave us all), but it’s certainly a sign of things to come. The more we hand over our intimate personal details and data to the machines, the more power they will have over our lives. Think of the Facebook image outage as the proverbial canary in a coal mine.
As one social media critic pointed out, it’s only when a system breaks down that we realize what’s happening at all. For more than 1.5 billion daily active users of Facebook, they came face-to-face with the AI algorithms powering their daily social media lives this summer. And what they found was not always pleasant – a clear warning to future Facebook users.