The Day AI Told Me I Couldn’t Talk to a Human
A glimpse into a future we should all be worried about.
Yesterday I wasn’t trying to start a movement. I was simply trying to fix a Facebook Page. I expected it to take fifteen minutes.Instead, I spent more than two hours trapped inside a support system that felt less like customer service and more like a maze designed by an algorithm. It wasn’t the bug that scared me. It was what the bug revealed.
The Conversation That Never Happened
Like millions of business owners, I pay Meta every month. Not because I love advertising. Because that’s where my customers are.
When something broke, I did what every customer has done for decades. I tried to contact support. Instead of reaching a person, I reached AI. The AI confidently instructed me to submit a bug report before a human could review my issue. That sounded reasonable. Except for one problem. There was no way to submit the bug report. I asked where. The AI crashed. I tried another route. The support page failed. No ticket. No email. No phone number. No human. Just another message telling me to “try again later.”
The AI confidently instructed me to submit a bug report before a human could review my issue. That sounded reasonable. Except for one problem. There was no way to submit the bug report. I asked where. The AI crashed. I tried another route. The support page failed. No ticket. No email. No phone number. No human. Just another message telling me to “try again later.”
The Scariest Part Wasn’t the Error
Software has bugs. Every engineer knows that.The terrifying part wasn’t that Meta had a bug.t was that the bug existed inside the only system capable of reporting bugs. Imagine a hospital where the emergency call button is broken…
…and the only way to report the broken emergency button is by pressing the emergency button.
That’s what happened. The support system became its own prison.
AI Didn’t Solve the Problem
It Protected the System
This wasn’t malicious. It was worse.
The AI wasn’t trying to help me. It was trying to complete its workflow. Its objective wasn’t: Solve the customer’s problem. Its objective was:
Move the customer through predefined steps. When reality no longer matched those steps….The conversation ended. Not because my problem was solved. Because the workflow had nowhere else to go.
This Isn’t About Facebook
Tomorrow it could be your:
- Bank
- Insurance company
- Hospital
- Government office
- School
- Payroll provider
As organizations replace people with AI, they’re quietly removing something we’ve depended on for generations:< The ability to say, “Can I please talk to someone?” That sentence has become surprisingly difficult.
Customer Service Used to Have an Escape Hatch
Every automated phone system eventually had a secret. Press zero. Say “representative.” Mash random buttons. Eventually…A human answered. Today’s AI often removes that escape hatch entirely. If the AI misunderstands you…There may be nobody left to correct it.
The Future I’m Worried About
This isn’t a story about Facebook. It’s a story about power. When AI becomes the gatekeeper between people…Who decides when you’re allowed to speak to another human? When software determines whether your problem deserves a person…Who audits the software? When the only way to challenge AI is through AI…Who do you appeal to?
Businesses Should Be Careful
I’m an advocate for AI. I build with AI. I use AI every day. But AI should make humans more accessible. Not impossible to reach.
The best customer experience isn’t one where AI replaces every employee. It’s one where AI handles routine questions quickly…and immediately recognizes when a person needs another person. That isn’t weakness. That’s good design.
This Is Bigger Than Meta
The problem isn’t artificial intelligence. The problem is designing systems with no exit. No escalation. No accountability. No human responsibility. No human oversight. No human responsibility. Technology should remove friction. Not remove people.
Read our open letter to Meta: Bring back the digital “Press 0” for business support →
One Question
If your customer service bank… your doctor…your child’s school…or your government office…adopted the same support model I experienced yesterday… Would you feel safer? Or would you feel trapped? Because after yesterday…I don’t think we’re arguing about customer service anymore. I think we’re deciding what kind of society we’re building.
Postscript
If you’ve ever found yourself trapped in an AI support loop—with no way to reach a human—I want to hear your story.
Not because I think AI is the enemy.
But because I believe the future should include a simple promise:
When technology fails, another human should always be reachable.