If you manage a Facebook group that’s been inexplicably banned over the last week or so, you’re not alone.
As reported by TechCrunch, thousands of Facebook groups have been suspended of late, with many pointing the finger at faulty AI-detection for incorrectly tagging them as violating the platform’s rules.
As per TechCrunch:
“Based on information shared by affected users, many of the suspended Facebook groups aren’t the type that would regularly face moderation concerns, as they focus on fairly innocuous content like savings tips or deals, parenting support, groups for dog or cat owners, gaming groups, Pokémon groups, groups for mechanical keyboard enthusiasts, and more. While some of the impacted groups are smaller in size, many are large, with tens of thousands, hundreds of thousands, or even millions of users.”
According to Facebook, the suspensions are based on a technical error, which is in the process of being rectified, with many group admins now being told that they should see their group restored again within 48 hours.
Meta provided this statement to SMT:
“We’re aware of a technical error that impacted some Facebook Groups. This has been resolved.”
So any errors should either be fixed already, or are in the process of being reversed.
But even so, the mass-bans have spooked a lot of Facebook group admins, who’ve spent years building their online communities.
As noted, broader discussion about the suspensions has pointed at erroneous flagging by AI tools, though neither Facebook nor Meta has confirmed that this is the root cause in this instance.
The concern here is that Meta is increasingly relying on AI, for an increasing array of functions, and that without human intervention, it runs the risk of seeing more errors of this type moving forward.
Indeed, Meta CEO Mark Zuckerberg recently noted that AI will replace most mid-level engineers at the company soon, and as it continues to expand its AI systems, it seems increasingly likely that Meta will be looking to AI processes to operate more elements of the business.
Which is why group admins are concerned about this latest wave of suspensions, as it could reflect a new change in approach at the app, where it becomes increasingly difficult to understand, and/or rectify problems, because you’re dealing with the decisions of the AI overlords, not humans.
Again, Meta hasn’t confirmed that AI is at fault in this instance, but it does seem like another example of Meta’s AI automation teething problems, which could point to further complications in future.
