With all the buzz about the AI-powered language modeling tools like ChatGPT, Replika, and ChatSonic, one thing is especially vexing me about this technology. (That is, beyond the sheer sci-fi shock that AI is now sophisticated enough to compose an intelligible narrative.)

I get that it's scary being one step closer to a Matrix-like robot uprising. I get that it poses significant security risks, making it super easy for criminals to impersonate people or create convincing malware. And we're all likely to be bombarded with even more relentless waves of manipulative digital propaganda and misinformation, for the rest of our lives, starting any minute. All this is duly concerning.

Load Error

But for selfish reasons, as an author, here's something that really ticks me off:

Sources are debating whether you should include an attribution line that acknowledges ChatGPT as the source of the articles it assembles. Meanwhile, the engine simply fishes for existing text and strings it back together in an allegedly meaningful way.

So who has actually written that text? The platform? The algorithms? Who should be credited as the originator of these pieces?

ME, that's who.

Me and people like me. ChatGPT will ride the backs of authors, marketers, bloggers, and copywriters who have been churning out essays, commentary, how-tos, wiki pages, and corporate website material for decades, in an attempt to educate (and, OK, even sell to) the public. As AI algorithms pore through all available online content, we now face the possibility of being the last generation of humans to compose original language for public consumption. Our words will be grifted, reinterpreted, and blithely reassembled into new iterations. (Ironically, AI is predictive-texting the end of my sentences as I write this...)

In effect, the intellectual property of copywriters, authors, and journalists who have been slaving away at their trades will be the source of all the articles that ChatGPT and its electronic cousins will spit out. From now until forever.

All I can say is, you're freaking welcome, everyone.

I once had a college roommate who was not known for her ability to ace essays. I'll call her Marlene. She majored in marketing and was rumored to be attending a major university more because her father was a wealthy professional than because she was a particularly dazzling student.

We were required to write a 10-page final paper. Marlene opened three marketing textbooks and pulled various paragraphs out of each one, flipping through chapters and grafting together passages, barely changing the wording. I don't know what grade Marlene got on that piece of work. But I doubt anyone who read it said, "Eureka! A brilliant piece of thought leadership!"

This is what ChatGPT sounds like to me. It's a plagiarism engine for the less-than-talented, even the ill-intentioned.

I understand this kind of offering could represent great savings for budget-conscious companies looking to produce basic website copy or "content marketing" blogs. As a freelance consultant, I routinely turn to company websites (and their competitors' sites) to get the gist of their missions. The thing is, I'm already shocked at how frequently I come away not having learned what those companies actually do. Frankly, many corporate website descriptions are nearly indistinguishable from one another, laden with jargon and generic superlatives. Show me a landing page that doesn't say "We listen to our customers. We solve your problems. We meet your needs," and I'll hire that copywriter on the spot.

If that's your requirement level, ChatGPT could be your best buddy. But be careful, because word on the street is that although these engines draft startlingly convincing language, they often get the details wrong. And do things like completely fabricate quotes.

The next question most writers ask is whether they'll soon be out of a job, the way factory workers were replaced by automation. My take is that the more people become dependent upon the flush of AI-generated text, the more they are lulled by the same regurgitated level of copy — the more prized the authentically unique, original thoughts of creative people will become in contrast. AI can generate language. But it can't create ideas (yet). And the more unresearched and pedestrian material these tools churn out, the more it will fill the sea of accessible content with generic and often erroneous, fabricated sentiments to be replicated again and again.

So for business people who are rubbing their hands together, thinking ChatGPT and similar tools might solve their content needs, a few words of advice:

Use the tools sparingly if you want to convey anything specific about your business that will truly differentiate you from the competition. Consider that your competitors have access to the exact same tools, and although content on the internet changes every moment, AI algorithms are all mining a near-identical pool of legacy source material at any given time.

Just because a language generator says someone made a statement doesn't mean they did. Investigate where the quote came from. If it appears in an existing article in the exact form you see it in your ChatGPT piece, then it's OK to use it in quotes — but you should credit the original author's article and the media in which it appeared. Otherwise, track down the speaker and run the statement past them for approval.

In defense of the creative mind, there's nothing as precious as an original idea. It's nice to let tools bulk up your web copy or compose your form letters. But you'll still need human insights to create something clever and strategic that will give your communications an edge, or an authentic and compelling voice to slice through the clutter.

I have faith in the power of creativity to overcome the robot uprising — the concept of which wouldn't exist without the fertile imaginations of talented writers, from Asimov to the Wachowskis. No matter what the tech industry comes up with next.

Start your unlimited Newsweek trial