Should we start talking about AI etiquette?
We probably all have that one friend or colleague with zero AI filter. They send around raw agent output: that all-too-familiar wall of convoluted text, full of phrases we recognise by now. The only appropriate response, obviously, is: “You are absolutely right. Thank you for your load-bearing message.”
The more tactful option would be to ask my own agent to summarise the message in human words and draft a response. The next logical step would be to remove the humans altogether and connect the agents directly. That doesn’t feel like an improvement to me.
Don’t get me wrong. I use AI for writing all the time, so I don’t think the answer is to avoid it. For now, I think a reasonable starting point is to read our own messages before sending them, make them easy for the other person to understand and take responsibility for the work our agents produce. “My agent did it” may explain how a mistake happened, but it doesn’t change who sent the message. That doesn’t mean I care whether it was written with AI. I care whether the sender has put some thought into the message and is willing to stand behind it.
Right now, I have to admit that I feel a little offended when I read obvious AI phrasing. I’m immediately more critical and less likely to read the message closely. Is that fair? Will we simply get better at using the tools until generic AI phrasing becomes less common? Or will obvious AI phrasing become the equivalent of sending a message full of spelling mistakes?