This morning, we published a post about a union member who used AI to write a five-part complaint about something that was actually good for him. A colleague read it and responded immediately:
"If some vendor or collaborator wrote an email to me using AI, I would go to moderate lengths to cut ties with them. At that point I don't trust their rigour."
Fair enough. Reasonable, even. If someone can't be bothered to write their own email, why would you trust their judgment on anything else?
But here's the catch. Two catches, actually.
Catch #1: You Won't Know
How would you be sure it's AI?
Right now, in early 2026, you can sometimes tell. The phrasing is a little too smooth. The structure is a little too clean. There's a whiff of corporate keynote in the word choices. But that's today. These models are trained on human text. They are stochastic parrots — and the thing about parrots is they get better at mimicking the more they hear.
AI has already been trained to sound like a person. It will be trained to sound like a specific person. To match your writing quirks. Your sentence length. Your habit of starting emails with "So" or ending them without a sign-off. The tells will disappear. They're already disappearing.
Your colleague sends you an email. It's well-written, clearly argued, and makes a reasonable point. Is it AI? Is it human? Is it a human who used AI to clean up their draft? Is it a human who dictated bullet points and let AI write the prose? Where's the line? When did it cross?
You won't know. And once you start wondering, you won't stop.
Catch #2: You Can't Always Walk Away
Our colleague's instinct is to cut ties. That's a luxury not everyone has.
A union has a duty of fair representation. When a member files a grievance — even a confused, AI-inflated, five-part grievance about something that's actually a win — the union has a legal obligation to respond. They can't say "we noticed you used ChatGPT, so we're not taking this seriously." The law doesn't care who wrote the words. It cares that the member raised a concern and the union addressed it.
The same is true in a hundred other contexts. A customer sends an AI-written complaint to your business — you still have to respond. A student submits an AI-written appeal to the dean — the university still has to process it. A constituent sends an AI-written letter to their representative — it still counts. A patient sends an AI-written description of symptoms to their doctor — the doctor still has to read it.
You can cut ties with a vendor. You can't cut ties with everyone. And AI just gave everyone the ability to produce professional-sounding documents that demand professional responses.
The Trust Problem
So here's where we are:
- If you detect AI in someone's communication, you lose trust in them.
- If you don't detect it, you might be acting on words that nobody actually thought.
- If you suspect it but can't prove it, you start doubting everything.
- And in many cases, it doesn't matter — you have to respond anyway.
Every outcome erodes trust. There is no path through this that doesn't.
This isn't about email. Email is just where you feel it first because it's personal and professional and daily. The same thing is happening to every channel: reviews, articles, comments, messages, reports, applications, testimonials. Every piece of text you encounter now carries a question it didn't carry two years ago: Did a person actually think this?
AI is eroding trust faster than anything humans have ever built. Not through malice. Not through deception, necessarily. Just by making it trivially easy to produce text that looks like someone thought carefully, when nobody thought at all.
The internet ran on trust. Not much of it, but some. Enough to do business, to collaborate, to communicate. You read an email and assumed a person wrote it. You read a review and assumed someone used the product. You read an article and assumed a journalist researched it.
Every one of those assumptions is now in question. Not because AI is malicious, but because it's available. The microplastics are in the water. The parrots are in the sky. And every email you open from now on — you'll wonder.