2026.23: Name Your AI and You Just Built the Perfect Scapegoat!
Happy Tuesday friends.
Sorry that I'm late getting this out...but I was enjoying some quality family time over the last few days.
Quick confession. I never call my AI by a name. I call it "my robot." Always have. My students think it's a bug. It's a choice. The second you give a tool a name and a personality, you start trusting it like a person, and you stop checking its work like a machine. So I keep it boring on purpose.
Turns out that boring little habit has a number attached to it now.
A Boston University researcher named Emma Wiles ran the same task past 1,261 managers. Same work, same output. The only thing she changed was the label. Half the managers were told the work came from a chatbot. The other half were told it came from an "AI employee" named "Alex", with a title and a spot on the team.
The group working with "Alex" caught 18% fewer errors.
Let's break it down.
Signal:
THE NAME DOES SOMETHING TO YOUR BRAIN BEFORE THE WORK EVEN STARTS
Eighteen percent fewer mistakes caught is not a rounding error. That is real money walking out the door in bad invoices, wrong numbers, and decisions made on garbage.
And it gets worse. When the tool was framed as an employee, two things happened to the humans around it. They felt less responsible for the output, like it belonged to Alex and not to them. And they were 44% more likely to bump questionable work up to their manager instead of just fixing it themselves.
Sit with that second one. You bought the AI to save your people time. The day you put it on the org chart, your people stopped trusting their own corrections and started forwarding its mess up the ladder. You didn't buy speed. You bought a longer queue.
Scale:
EXPERIENCE IS THE NEW BAR
That breaks something we never thought to question.
The apprenticeship is gone. For a hundred years, you learned by doing the grunt work. The junior analyst built the boring model. The young lawyer read the boring contracts. That's where taste came from. Reps, on small stakes, for years. The machine does the grunt work now. So how does anyone build judgment without the climb that used to build it? I don't have a clean answer. Neither does anyone else. But pretending the old ladder still works is how we end up with a generation that can prompt and cannot think.
Here's the part nobody wants to hear. Your AI agrees with you. About everything. It's a brilliant, tireless yes-man, and that is its most dangerous feature, not its best one. So argue with it. Make it take the other side. Ask it to tear your idea apart before your client does. Treat its agreement as a smell, not a reward. I don't want to be praised. I never have. I want to be engaged. Point your tools the same way.
And underneath all of it, your voice matters more now, not less. When every post sounds like it rolled off the same assembly line, the human one stands out like a sore thumb, in the best way. Your voice is built from your experience, your values, the things you actually believe at 2am. That's not a feature you can install. It's the one thing that stays scarce while everything else gets cheap.
THIS IS A BRANDING CHOICE, NOT A TECHNOLOGY CHOICE
Here is the part that made me laugh out loud. The article I'm pulling this from, is a sharp piece in MIT Technology Review. The paid sponsor sitting right next to it: KPMG. And in their ad, KPMG did the exact thing the article spends two pages warning you not to do. The headline reads: "The digital teammate era is here, leaders must act now." It sells digital teammates that "make decisions" and "take action."
So one column warns you that calling AI a teammate makes your staff 18% worse at spotting errors. The column beside it sells you the teammate. Same page. You cannot make this up.
I'm not picking on one firm. Every big vendor is doing it. Since spring, Microsoft, OpenAI, Anthropic, and Google have all shipped tools you "manage" like a team of little employees. Almost a third of the managers in Wiles's study said their company already frames AI agents as staff. Nearly a quarter put them on the actual org chart.
None of that makes the tool better at the job. It's a costume. And the research shows the costume makes the real humans worse, the same humans you're paying for the judgment the AI is trying to copy.
NAME IT A TOOL, KEEP THE ACCOUNTABILITY WHERE IT BELONGS
I've said it before in these letters: the machine lifts, you lead. AI can compress the time a task takes. It cannot hold the responsibility for getting it wrong. The moment your team believes "Alex" owns the output, nobody owns it, and a tool becomes a convenient place to dump the blame when a human decision goes sideways.
A Nobel-winning economist at MIT, Daron Acemoglu, put the alternative plainly. AI built to replace people is a losing bet. AI built to make people better at what they already do is the whole game. A tool in a capable person's hands, not a fake colleague you hope will cover for you.
So keep it boring. Keep it a tool. Keep your name off the org chart and your team's name on the work.
Deep Dive:
This week, do two small things.
One, go listen to how your team talks about AI. If you hear "ask Alex" or "the bot decided," that's not cute, that's the 18% leaking out. Rename it. Make it a tool again, out loud, in the next meeting.
Two, write down one rule before you expand any AI workflow: a named human signs off on every output that touches a customer or your books. Not Alex. A person, with a pulse and a paycheque, who owns it.
Thanks for reading!
Hit reply and tell me what your team calls your AI. If it has a name and a personality, I'd love to know whether you picked it or the vendor did.
The newsletter isn't the conversation. The conversation is the conversation.
See you next Friday.
Best, JT
PS - If someone forwarded this to you and you want it in your inbox directly, subscribe HERE.
