How Not To Write Like An AI

It will be a relief once ChatGPT and its ilk get going in Facebook groups. Not because it can be rude to people (there are already humans who are good at that), but because it can answer recurring questions like “what is the weather” “how should I pack” and “which part of Tokyo is better”.

Then, those of us who are interested in sharing our experiences and tips can focus on people who really need answers.

It will also address the strange entitlement that you often notice in Facebook groups, that people somehow feel they have a right to get an answer, even though the groups are run on a volunteer basis. Perhaps Facebook will start sharing some of the income with the moderators who are able to maximize ChatGPT use.

But while Facebook has almost killed off all discussion services outside itself, it is not invulnerable. And someone who is out to grab a piece of the pie Elon Musk has dropped will accidentally come up with something better. Probably including AI in some way.

What AI can do is answer questions based on available information in a slightly nicer way than a Google search. But apart from when it makes up information to sound credible, it can not create new information. Which means that a clever human can find it first, usually in even better representations than pop up on the Google front page.

But an AI can not answer the question “what did it feel like” unless someone has put the answer out there. That is where the creative part of the web (and writing in general) will go. Youtube will take a few years more, but someone will figure out how to make an interesting film from Streetview, and then Youtube will be as boring as the web today.

Published by

Leave a comment