Is Generative AI the Death Knell for Tech Writing
I had a wild hair a while ago to take this massively open online course in Computer Science. There was a series of recorded video lectures and I’ll never forget one of them, because it was still the best explanation I’ve witnessed to date for how computers work. The professor had a couple of volunteers come to the front of the class and give instructions to the TA (who was acting as the computer) on how to make a peanut butter and jelly sandwich. One student confidently took the mic and gave the first instruction: “Open the bag of bread.” And the TA, big mischievous smile on his face, stabbed the bag of bread open with a knife repeatedly. Bread flew everywhere. In that moment I understood two things:
- I would’ve gotten a kick out of being that TA and coming up with increasingly idiotic ways of doing simple tasks students asked of me.
- Computers only do exactly what you tell them to do. No more, no less. They need parameters and specifics and details.
The second thing I assume was the lesson. Point taken. Lesson learned. 10/10 for the visual aid. Thank you massively open online course that I did not finish.
Are we in Westworld or Westeros
Lately I’ve witnessed just the biggest of big daddy freak outs over the introduction and mass adoption of AI tools like ChatGPT in the technical writing community. People are wondering what it’s going to do to the tech writing profession. The hot takes seem to be a mixed bag right now. There’s a lot of confusion about what it is, what it’s capable of, and how much like Skynet it will become.
And it’s not just the technical writing community. People are generally like “oh no my job” right now everywhere, especially if you have writer in your title, because one of the things these LLMs are getting pretty good at is language and writing. They are Large Language Models. So naturally, you would tend to think their language prowess is large. This disruptive technology is at least one reason why the WGA and SAG is on strike and some of my favorite shows are delayed (come back HOTD I need your platinum blonde weirdness and negroni sbagliatto with prosecco Tik Tok moments). Doing a Google around AI and writing yields many anxious headlines attached to anxiously written opinions usually with the take: no, no it’s not going to kill all the writing jobs. At least not right away.
From what I’ve gathered, there are a couple schools of thought. Some folks think AI is going to have us all busking for bread on the side of the highway, while others are saying it will help us become more efficient and will make daily tasks less tedious and cumbersome.
I tend to be in the latter camp.
I’m a slow adopter of things in my home life. For example, it is 2023 and I just now bought an Alexa-enabled device for home automation on Prime Day, however, I will admit to you here and now dear reader that it’s still in the packaging and I dooooo not know when I shall be removing it from said packaging and doing anything with it? But when it comes to work stuff, I like to be on the bleeding edge of what is cool, which has changed over the years (and yet I still somehow can’t escape MS Word).
Right now, AI is pretty dang cool. If Twitter was working appropriately and not on the verge of a rebrand/collapse at the hands of one of our most ridiculous billionaires, AI would be a continually trending topic amongst the twits.
The (job) safety dance
I sometimes go give little talks to college students, and the last time I did, one of the questions they asked me was if I thought AI was coming for their jobs. And the answer I gave them at the time was something like, “Nah, don’t be worried, needs more iterations.” And while that’s probably true for now, as things improve over time, I think it’s probably likely that AI will have some outside impact on the amount of open headcount and the amount of redundancies in the tech writing field across the board. Especially if you’re in the software space, which is where I tend to lurk. As one friend of mine at a different software company put it in a text to me that I bet he didn’t realize I was going to use for a Medium blog, “yeah, it’s totally coming for their job from an executive perspective.” And in business, unfortunately the executive perspective is a thing you kind of have to consider.
Am I worried for my job? Not really. I know I’m pretty smart. I’ll adapt. I got a mortgage, son. I need to stay employed.
Am I worried for these kids coming up into the workforce? Also no. I know they’re pretty smart and they’ll adapt.
You can absolutely control what you fear, or whatever Winston Churchill said. AI is a tool. Treat it like a tool, learn how to use it, and you won’t have a reason to fear it. (Until Skynet, eventual societal collapse, etc., but that’s another post.)
*yells dramatically* Alas, what is to become of tech comm?
In Tom Johnson’s blog I’d Rather Be Writing, he mentions how little technical writing technical writers actually do, which I and the lol-worthy number of tickets assigned to me find to be excruciatingly accurate. In The influence of language-generative AI tools on tech comm: parlor tricks or disruption? he points out:
Perhaps the only saving grace is that technical writers spend only a fraction of their time writing documentation… What perhaps AI can do is take existing engineering babble and convert it into more readable content.
That’s absolutely something that I think will come of this generative AI period in time. Instead of using technical writers as filters and content bar raisers, engineers can talk directly to the AI chatbot, and the AI can translate. This is the type of delegation we need and deserve. You translate the engineer-speak, lil AI, you do it. Imma get me a Kit-Kat and a coffee, you won’t find me desk-side.
AI can be leveraged for other things too, like grammar updates, rewording and improving readability, matching style, legal, marketing, and security requirements — the minutiae of tech writing work. Particularly because I don’t really love doing that stuff, and as we all know, the world revolves around me. There are a metric ton of new AI tools to play with in the technical writing space, and I’m sure more will be coming out, probably to solve very specific use cases, and none of which will be approved by your security team, but you know, it’s fun to imagine things that could improve your life actually being approved by AppSec for you to use. So there’s that. It’s the dream. Live inside the dream. Don’t come out. It’s weird here.
In this Survey of Technical Writers on AI there’s a lot of interesting data presented, but some of the biggest takeaways are that tech comm people are mildly concerned about the injection of AI into their lives, and people think that tech writing as a profession will skew more toward content strategy or information architecture. I don’t disagree. And in fact, AI overlords, if you’re listening, please skew my job that way. I’ve been hunting for comma splices long enough. Put me out to pasture, but let me play with the big problems — there’s too much content, people can’t find what they’re looking for, stuff gets outdated quickly, and also there’s not enough of the right content, but what that is is a mystery because feedback mechanisms haven’t changed in like 20 years.
Look, my take is no different than anyone else’s right now. Mostly because AI hasn’t done anything scary or alarming yet. It’s just out there being the cool new hot toy.
Do I think it’s coming for tech writing jobs? Eh. In a few years maybe there will be fewer jobs for tech writers. The jobs will probably morph and become something else, but that’s okay. Change is constant. The death of one thing is usually the birth of another. *Something something third generic platitude.*
In its current form, the more I work with and use generative AI, the more I realize how much it needs people. It requires input to be effective. If the input is incorrect, or vague, then it will be incorrect and vague in response, just like anything else computer-related. I hope you just pictured the bag of bread exploding. I know I did. And that gives me hope for people that like to poke holes in things.
Jen Jarnefeldt currently does technical writing for AWS, and she’s been picking apart what other people say for her entire life. Jen loves ice cream, fat babies, meme culture, and rock climbing. She also remains pretty sure that Elon Musk is an alien.