Sorry, who invited the AI?
Why we need a ‘no AI without consent’ rule for workplace meetings
A few seconds into a recent call, I clocked a participant I didn’t recognise. No name. No video. Just: Otter.ai is recording.
Someone had brought an AI note-taker to the meeting. No one had mentioned it.
And no one else seemed fazed either. I got the distinct sense that I would be the awkward one if I said something (which I did, because I am okay being awkward).
And that’s a problem. AI agents are sliding into our working lives without so much as an introduction. They arrive unannounced, remember everything, never get tired or distracted. And yet, somehow, it’s considered impolite to ask them to leave.
We need to change that.
“It’s just a transcript” is not a neutral phrase
The standard defence is: “It’s just for notes.”
But tools like Otter, Fireflies, Zoom AI Companion do far more than jot things down. They summarise conversations, extract action points, analyse tone, flag key moments, sync to CRMs and cloud drives - and in some cases, train future models.
We’ve gone from helpful notetaker to always-on observer - without ever deciding if that’s what we want.
And no one is ever asked to consent. Even if you’re fine with your own AI tools, you might not be OK with someone else’s AI quietly capturing your voice, your ideas, your data and a perfect image of your home office.
When there’s AI in the room, everything shifts
While many people don’t seem to take much notice, for me, the dynamic changes. People hold back. They choose words more carefully. Candour drops.
That’s not paranoia. It’s psychology.
Being recorded changes how we behave. The social contract of a meeting - live, collaborative, human - starts to unravel.
Brainstorming becomes performance. Negotiation becomes theatre. Trust erodes.
The etiquette void
Why don’t people speak up?
Maybe because it feels awkward. Because the meeting’s already started. Because no one else is objecting. Because we haven’t yet decided if asking a bot to leave is rude.
This is how surveillance creep works. Boundaries erode slowly, justified by convenience and tolerated through silence.
This is a decision point
The AI tools being normalised now will define what’s “standard” later.
Which is why we need a clear norm:
No AI in meetings without informed consent from everyone involved.
Not a vague “recording for notes” buried in the invite. A proper heads-up. With a real option to say no.
If you’re not sure how to raise it, try:
“Just flagging - do we all know there’s an AI recorder on this call?”
“I’d prefer to go off-record - can we turn off the bot?”
“Would everyone be OK pausing transcription? I want to speak freely.”
Not dramatic. Not disruptive. Just normalising consent.
For businesses, this is bigger than etiquette
If your company is using AI to record meetings - internally or with clients - you need a policy. Not just for compliance. For culture.
Because here’s what’s coming:
Confidential discussions leaked
Legal risk from undisclosed recordings
Employees self-censoring when AI is “listening”
If AI note-takers are quietly becoming part of your team, start treating them like actual team members: onboard them properly, declare them clearly, give others the right to opt out.
There’s a reason we don’t default to recording in-person meetings. We instinctively know it changes things.
Same applies online.
Just because you can bring AI to every meeting doesn’t mean you should.
Sometimes the most human move is to hit pause - and ask who (or what) is actually in the room.
