I’d say probably the biggest one would be data entry/billing jobs.

Mainly, I take issue with the early adoption of these services. I’m all for using GPTs and image generators to generate boilerplate material that people can then review and save time with. It seems like a lot of the companies trying to use these services are looking to replace jobs that really need a human touch.

In this case, grading papers isn’t just to tell the student where they can improve, but the teacher as well. If the teacher sees a trend in a common failure point, they know they need to work on the lesson plan for that topic.

It also rubs me the wrong way that the chat services don’t have any qualifications to meet on top of the fact that they’re constantly changing and often have “bad days” where some things get configured wrong.

If we bring tools like this into something as critical as education, what happens when the workload that they’ve been used to covering with a GPT has issues for several days? It should be stable before you bring it to the masses.

At least right now, we should stick to using it for trivial things and occasionally assist with others. The idea of handing material over to a GPT (not even getting into many potential privacy concerns) and trusting it? I just don’t think we’re there yet.

You can undo or reverse a transaction in a billing software. You can roll back code changes, etc. but you can’t fix an education after the fact. That becomes a responsibility of the student which is unfair.