Maybe it’ll stick


Maybe it’ll stick

In the 1984 movie, “Terminator,” we were reminded that Skynet was coming. Artificial intelligence would rule the world. Someday. Yes, maybe today.

OpenAI introduced ChatGPT 3.5 (whatever happened to versions 1, 2 and 3? I wonder), a generative AI chatbot that combs all sorts of online information sources to provide text response to questions. ChatGPT can write legal briefs, essays and term papers. For free! (It’ll cost you $20 a month for Version 4.)

Does it work?

Are the results enough to shake a stick at?

I think chatbot AIs are risky, dangerous and evil.

Human laziness is incalculable: it has no end. We will always do our best to avoid work: sometimes, for some of us. But for those who can write fiction and great, well-researched and documented reports, oh please continue to do so.

Don’t undervalue what you do. Professional, HUMAN writers are going to bring a level of experience, creativity and insight that a chatbot text generator just doesn’t have. You have that wisdom that the still-immature, flawed AI doesn’t have.

ChatGPT 4 is supposed to be more advanced. But when it is tested, the chatbot sometimes makes up facts and creates faux cases. In one case filed in February 2022, in an aviation lawsuit, an attorney was fined $5,000 in June 2023 by a U.S. district judge when it was discovered the facts were not checked correctly under due process. So, yes, you’ll have to review the generated text.

I think it’s hard to beat us HUMAN writers. But I’m really afraid journalists will start depending on ChatGPT. Newsrooms will do away with entry-level staff to save money. Facts won’t get checked. Publishers won’t get it right: They’ll get it by deadline. Throw it on the wall; maybe it’ll stick. Maybe it will adhere.

Or maybe it will just slide off, gross and slimey.

Signature_Andrews-Andrew