GenAI has been available to the public for, what, a year and a half? And now that we’ve settled into it, I would say there are two kinds of people who lean in to using GenAI:
ONE: People who are not familiar with the subject matter in question.
If you do not know JavaScript, ChatGPT’s JavaScript looks like wizardry. If you do know JavaScript, it can look like spaghetti. Likewise, if you are not a writer (or, let’s be frank, a reader), Claude seems brilliant. If you are, you’ve seen one Claude paragraph, you’ve seen ’em all.
I’ve seen multiple demos of GenAI translation features where a person who does not know one of the languages in the language pairs gets some output and says “wow, that’s amazing!” with absolutely no way of knowing whether the translation was accurate or even vaguely correct.
Subject matter noobs don’t know what they don’t know, and GenAI seems confident, so it easily draws in the naïve.
TWO: People who are familiar with the subject matter in question, but don’t care if the output is mediocre or incomplete
Let’s be honest, many people cut corners, and GenAI is the perfect corner-cutting tool. They don’t care if their e-mails sound like they came out of a can, they don’t care if the memo summary is incomplete, they don’t care if the people in the image have extra fingers, etc.
The genuinely sloppy users of GenAI are what they are. Some people aren’t paid for quality work, and so they don’t produce quality work. We’ve all seen the campaign ads where the faces of the people in the background are melting. Or the outright scams like fake e-books and SEO spam. Fine. These people will always exist, this stuff will always be around.
More concerning is professionals who get GenAI output that sounds plausible and say “good enough” and call it a day, without even checking to see if this frequently-wrong technology in fact performed the task correctly.
I’m thinking about doctors who use GenAI to transcribe and then summarize conversations with patients; standards compliance reviewers who use it to check large PDFs against requirements; lawyers who use it to write contracts; technical writers who use it to write white papers. Lots of documents aren’t really produced in order to be read: They are produced in order to check a box, even though if they are actually needed someday, it is crucial that they be correct and complete.
But GenAI is a technology that makes it very easy for people working in areas with poor visibility to produce deliverables that look plausible and complete, and without a detailed review, no one will know any different.
My suspicion is that this will eventually dawn on management and we will see corporations implementing policies to restrict or at least monitor how GenAI is used in the workplace. Eventually, GenAI use is going to be a scarlet letter, an indication that a product, service, or professional is low-effort and poor quality.
Unfortunately, it could be some time before this happens, and in the meantime, mediocrity accrues.
Leave a Reply