According to engadget, Meta is launching a new AI-powered effort to fix its notoriously broken customer support system. The company is rolling out a new “support hub” on Facebook and Instagram that centralizes help features and includes an AI chat for questions. It’s also testing a new “AI support assistant” on Facebook for personalized help with issues like account recovery. For hacked accounts, Meta says it’s improved the process with better alerts and new verification methods, like an optional selfie video. The company claims these AI-driven changes have already increased the relative success rate of hacked account recovery by more than 30% in the US and Canada this year. The most reliable way to get live human support currently remains a paid Meta Verified subscription.
The AI Hail Mary
Look, this is a classic case of a company trying to automate its way out of a human problem. Meta’s support has been a black hole for years, a fact they’re now openly admitting. Throwing AI at it feels like the obvious, maybe the only, move for a company of its scale. The promise of “instant, personalized help” sounds great. But here’s the thing: the real test isn’t handling simple password resets. It’s the weird, complex, edge-case problems where automated systems completely fall apart. You know, the exact situations where people get desperate enough to sue the company in small claims court. An AI chatbot repeating policy links won’t cut it there.
The Access Problem
And there’s a huge, glaring flaw in the plan. The new in-app support hub is useless if you can’t get into your app because you’re, you know, hacked or locked out. Meta points to its external account recovery tool as the solution, but that’s often the very system people complain about. The improvements—better alerts, recognizing your familiar devices, the selfie video option—are genuinely good steps. They might actually move the needle on that 30% success rate they’re touting. But it’s starting from such a low baseline that a 30% improvement might just mean going from “impossible” to “merely incredibly difficult.”
The Human Touch Still Costs
So where does this leave us? Basically, Meta is creating a multi-tiered support universe. For the masses, it’s AI chatbots and automated flows. For the slightly more serious user willing to pay, there’s Meta Verified and its chat-based support (which, let’s be honest, many say is still limited). The unspoken message is clear: getting a real human to thoughtfully fix a unique problem on a free product is becoming a premium luxury. The AI assistant, as detailed in their official update, is in early testing. The big question they haven’t answered: will it ever be smart enough to escalate to a real person when it’s stumped? Or will it just hit a dead end with a polite, “Sorry, I can’t help with that”?
A Step, But Not A Solution
I think we should be cautiously optimistic about the specific account recovery tweaks. Using AI to recognize your typical login patterns is a smart, non-intrusive security measure. But let’s not confuse better automated tools with actual customer *service*. Meta is optimizing for scale and cost reduction, not for user empathy or complex problem-solving. They’re treating support like an engineering problem to be solved with smarter algorithms. Sometimes, though, people just need another person to listen and take ownership. Until Meta builds a bridge between its AI and a competent, empowered human team, the core frustration will remain. It’s a step in a better direction, but the destination still seems awfully far away.
