Skip to content
Liene Daktina

Another Way to Look at Policy

We buy AI tools from suppliers, what are we responsible for?

01 AI ETIQUETTE

Five AI etiquette breaches nobody has written down, and everybody notices

What fixes them, and where the EU AI Act stops.

Liene Daktina · 28 October 2026 · 5 min read

Most firms have a rule for phones in meetings and none for the assistant that now writes the minutes. So the rules get made in the thick of things, by whoever happens to be holding the phone. A notetaker joins a call unannounced. A summary goes to a client with the model’s mistake left in. A colleague receives a reply that was plainly generated and reads it as the time you did not have for them. Nobody complains. Everybody has, at least once this year, felt mildly outraged at the sheer temerity of others, and said nothing.

Office etiquette used to be short and settled, and most of it was never written down because it did not need to be. Do not take credit for another’s work. Do not waste the hour of the person across the table. If you want something read, write it. Thank whoever made your day easier. The assistant has not changed those rules. It has made each of them cheap to break, hard to notice, and harder to call out, because the breach now arrives switched on by default. Meeting notes write themselves unless somebody turned them off. A new channel comes with an agent already seated in it. The mail client finishes the sentence before the writer has decided how it ends. Much of what colleagues hold against each other was never chosen by a colleague; it shipped on, and the firm has not yet made the decision it believes it made.

What clients tell me in the sessions sits close to the research, and the two paint the same picture. Managers, lawyers, the occasional managing director: they name the same five habits the surveys find, in nearly the same order.

  1. Generated work forwarded unread, with the errors left in. Three in five workers call it the most irritating habit in the office.

or

New perspectives on European innovation, policy and responsible AI.
No spam. Unsubscribe anytime.