Manners for the AI era.
In 1994, Virginia Shea wrote Netiquette, a guide to behaving decently on a strange new medium called the internet. Her first rule was “remember the human.” AITiquette is that same principle, but in the context of the strange new(ish) world of GenAI. A few of these considerations echo responsible AI; however, the distinction is that this protects the person your work reaches, while responsible AI protects the author from consequence.The model can produce the words. It cannot stand behind them, and the person on the receiving end still needs someone who will.
Considerations
Read it before you send it.
AI gave you a draft. What you send is yours. Read it, fix what's off, and make it something you'd stand behind.
Spend the time you save improving the work.
AI can hand back 45 minutes. Put them into sharper work, not into looking like you spent the hour. Faster is only a win if the output gets better too.
Community suggestions
What to do with the time is a really tough question because I think people would just use it to goof off. I’m not sure of the answer, but I do think this needs more strength and a clear tactic. Sharper work makes sense, but I’m not sure people will know what to do.
Be upfront that you used AI.
If someone would want to know AI was involved, tell them. People can handle knowing how the work gets made.
Community suggestions
Everyone will use AI for almost everything in the future. Not a key point imo as this will be a given.
Question is, how do you ask someone if they want to know? Maybe you put that as a tag in the same way that people tag this is sponsored content.
Keep confidential information out of public AI.
Client names, a colleague's salary, a friend's diagnosis. None of that is yours to feed to a public model, and the person it belongs to never agreed to it. The point is not the policy. It is the trust.
Fact-check what AI gives you.
AI invents statistics confidently. Before that number goes in your deck, check it. The hallucination is AI's bug. Using it unverified is yours.
Community suggestions
Not just statistics, but other facts. Might consider changing the word ‘deck’. Might not be fully understood in same way across all industries.
Don't let AI flatten your voice.
AI will write in your style and quietly sand off whatever makes it yours. Your specific take is the reason people read you, not the prose. Don't outsource it.
Community suggestions
I’d strengthen this by saying, Don’t lose the skill of writing, because it’s going to be so helpful. No one is going to write anymore, so if you can still write, that’s a massive advantage.”
Ask before AI records a conversation.
Before spinning up an AI note-taker, check with everyone else in the room. Most won't mind. But it is not your call to make alone.
Community suggestions
That's very specific to virtual business meetings. Suggest rephrazing it to reflect something more general.
Write less, not more.
AI makes volume easy. Your reader's time is not unlimited. A long AI-padded document is not more thorough than a short direct one. It is just harder to get through.
Community suggestions
This is what we're all learning: everyone's generating huge, big AI decks, and then no one reads them.
Don't take full credit for AI's work.
If the draft or the structure came from a prompt, say so when it matters, especially when others assume you did it alone. Using the tool is fine. Quietly claiming all the effort is not.
Help others get good at this, too.
Not everyone is at the same point with this. If you are ahead, share what you know. AI fluency is not a competitive edge to sit on. Pass it on.
Proposed considerations
Community-submitted ideas reviewed by the editors. Not yet part of the official list unless promoted there.
All suggestions are pretty much targeted at someone who uses an LLM as a google replacement of sorts. Is there room for other items that go beyond?
Could consider adding some themes to group these: Onwership, respect, improving your work - or something