There’s a guy named John at the ad agency I work for. John is one of my “associate copy chiefs.” I’ve worked with him for almost five years now across three different companies. (I just keep hiring him wherever I go.)
John is also quite possibly the biggest proponent of A.I. at our agency. The things he does with it veers on David Blaine-level magic. He’s trained a million bots (A.I. tools) to research, come up with hooks, mechanisms, and headlines… and even write pretty decent copy that converts.
But every time I get him to train the rest of the team on what he does, he will always start the lesson with:
“A.I. is stupid and it will lie to you.”
I think a lot of writers right now are justifiably scared of A.I. destroying their livelihood. I’ve seen what it can do. Our agency leans heavily on its use. And if you’ve spent your entire career writing SEO slop “content” to feed the algorithms… I think you should be scared. A.I. will eat your lunch.
However. If the kind of writing you do requires what I’m looking for in a copywriter (everything I talked about in parts two and three of this series)… then I don’t think you need to be frightened).
Why? Because “A.I. is stupid and it will lie to you.”
Let’s talk about what A.I. is. First, it’s not actually A.I. At least not the A.I. science fiction promised us. It’s not sentient and it never will be. It doesn’t actually “think.”
The A.I. we got in this universe is a “probability engine” that’s based on a compressed “blurry image” of data.
When you prompt A.I., it spits out a word, then based on probabilities of what the next word is, it will spit out the next word, then the next one, and so on. Those probabilities are based on all the words we threw up on the Internet since 1995 (or whenever AOL was sending us a gajillion CD coasters).
Think about that for a second. Think about who posts the majority of content on the Internet.
That’s not INTELLIGENCE. That’s just a really fancy search engine regurgitating words remixed to answer your prompt.
In order to write intelligently, you would need to give the topic CONTEXT. You need to DISCERN what’s important, what’s not, why, and how it “fits” with your overall argument. You need to create FRAMEWORKS on how the audience should think about what you’re talking about. You need to explain complexity with clear LOGIC CHAINS and fresh ANALOGIES.
A.I. cannot do that.
Now, can you prompt A.I. to do that? Yes, but… “A.I. is stupid and it will lie to you.”
Think of A.I. like this. They’re basically a precocious ten year old kid who can calculate and recall information really fast based on everything that’s ever been written online. Add on the fact that it will say things to make you happy, including lies and hallucinations. (Most kids want to impress their parents.)
So in order to use A.I. the way John uses A.I., what do you think is required?
One, that you’re smart enough to tell when it’s acting stupid, and two you’re always skeptical of the answers it gives you and recognize when it’s lying.
John will spend anywhere from 2 to 6 hours building out a single bot. And even then, he’s constantly stress-testing them to ensure it does what he wants them to do. He is essentially training the bot to think like him. But like a bad employee, you still have to micromanage them.
In other words, the only people using A.I. well are people who can already do said thing well. At our agency, we are extremely careful with letting junior copywriters use A.I. They simply don’t have the experience, wisdom, and expertise to recognize when “A.I. is stupid and is lying to you.”
So with that in mind – my final message to you for this series is, yes, you can still be hired as a copywriter and get paid really well in the age of A.I.… provided you BECOME a writer that does what only humans can do.
I believe you can do that. I know that if you lean in on mastering your craft, you’ll succeed. And I look forward to meeting you and quite possibly hiring you when you’re ready.