I’m not a massive fan of social media. The only platform I use is LinkedIn—and even then, I often find myself questioning the life choices that led me to need it. This will come as no surprise to anyone who knows me.

I mention this because I need a surfboard and, in my area, the secondhand goods market has largely moved to Facebook Marketplace.

So, fifteen years after committing Facebook suicide, I decided to open a new account.

My account was active for exactly three days before being banned for life for “violating Community Standards.” The decision is final and cannot be reviewed or discussed. This is the message I now get when trying to log in:

The sequence of events that got me here isn’t hugely important. Still, I’ll sketch them out in case my conclusion is incorrect and something else is at play.

To open the account, I provided only an email address and a birthdate. Full disclosure: the birthdate wasn’t accurate. I assumed Facebook had some minimum age requirement, but beyond selecting the correct year, I didn’t think they needed precise information. Once the account was created, I was guided through the onboarding process: upload a photo, add friends, and so on. I skipped all of that—I just wanted to get to Marketplace.

And honestly? Marketplace was intriguing. Despite my jaded cynicism, I quickly fell down a rabbit hole. “Antique” wood planes, knife sharpeners, even a few houses—I browsed items I had no intention of buying (except maybe the knife sharpener, more on that in a moment).

The next day, I decided to actually search for surfboards—the reason I joined in the first place. But when I logged in, I was told my account had been suspended pending verification. Facebook asked me to provide a mobile number. Once I did, they sent me a PIN via WhatsApp.

At this point, I was getting irritated. Why did Facebook need my phone number? Was this just profiling? Trying to be less cynical, I considered that perhaps they were attempting to verify users and weed out bots. And fair enough—my phone number is 25 years old, and my WhatsApp account nearly as old. Surely, this should prove I'm human.

Apparently, it did. My account was reinstated, and I was back on Marketplace.

I didn’t find the board I wanted, but Marketplace still seemed promising. I figured I’d check in periodically and maybe set up notifications. Just before closing Facebook, I remembered the knife sharpener. Knives should be sharp. They work better that way. I messaged the seller to check if it was still available and to ask a few questions that weren’t listed. Then I closed Facebook and got on with my life.

On the second morning, I received an email from Facebook:

That was... unexpected. I logged in and was told I had to upload a video of myself for verification.

Now I was annoyed.

Why would they need a video? This wasn’t about verifying that I was human—any basic AI could use metadata from my WhatsApp account to determine that. That account is old and well-used. And yes, I’m fully convinced they do use metadata, if not the encrypted messages themselves. This didn’t feel like security. It felt like data collection, plain and simple.

After a one-person rant, I decided to comply—I still wanted that knife sharpener. I recorded a video of my face, turning left and right as prompted. I was told it might take up to 72 hours to verify and that, in the meantime, my account would remain suspended.

A few hours later, I got the final email: my account had been permanently disabled. No explanation. No recourse.

I did some digging. My findings aren’t 100% conclusive, but everything suggests that an AI system reviewed my video and flagged me as... not human.

I’m not sure whether to be flattered or insulted. (Have you noticed AI-generated images always seem to be of good-looking people? Perhaps I'm just so good looking there is no way I could be real)

But mostly, I felt furious.

Here is a company—one that positions itself as a global AI leader—and they can’t even build a reliable classification model. Worse still, they offer no appeal process, no contact point, no escalation. Just silence.

And now, because their system flagged my phone number and email as suspicious, I’ll likely never be able to open another account with either of them—or anything linked to them.

Yes, I’m disappointed I can’t access Marketplace. But what really frustrates me is this: Facebook—supposedly a world-class AI company—seems to have forgotten a fundamental truth about AI systems: they are statistical tools. They will never work with 100% accuracy. They require error handling. They need human oversight.

Failing to acknowledge this shows a profound misunderstanding of the very technology they claim to champion.

Of course, Facebook can afford this. They’re a monopoly. Whether I use their platform or not makes no difference to them.

But if you work in a space where you don’t have that luxury, take note: poor implementation of AI—especially AI without a human safety net—will drive your users straight to your competitors.

Rant over.

Back to work.

But do try to remember: your AI solution needs a human in the loop...
Unless you work at Facebook.