The Future of Technical Writing
What I read: The College Essay Is Dead by Stephen Marche. Published December 6, 2022
When I read the article referenced here, I was somewhat startled. I’ve followed the developments in automation, artificial intelligence (AI), and robotics closely. These topics fascinate me. I’m well informed, far more than most. Even so, I was surprised at how advanced AI had already become.
Pass or fail? A- or B+? And how would your grade change if you knew a human student hadn’t written it at all? Because Mike Sharples, a professor in the U.K., used GPT-3, a large language model from OpenAI that automatically generates text from a prompt, to write it. (The whole essay, which Sharples considered graduate-level, is available, complete with references, here.) Personally, I lean toward a B+. The passage reads like filler, but so do most student essays.
Recently, OpenAI released it’s ChatGPT technology to the public. A lot has been written about this, this AI is finally good at stuff, and that’s a problem article as but one example.
In some cases, alarm bells are going off in people’s minds about what AI is already capable of and the news and technology commentary cycle and their reading public have taken notice. They should. AI is on the precipice of upending everything, and I mean everything.
This isn’t news. Much has been written about AI’s impact on our future. But the sudden realization among the masses that AI isn’t just a future thing, but is here now and poised to change our lives forever, was hammered home when ChatGPT was released with such remarkable functionality demonstrated even in its early iteration.
Reading the article made me reflect on what such advances might do to my former career as a technical writer. Until September of this year, I was a Senior Director of an Information Development department at a Fortune 100 enterprise software company. At various times I oversaw technical writing, curriculum development, video tutorial development, and other information-related projects that supported a popular application product line.
Adding together all of my years working directly with computers, including my time in the technical information realm, I have about 34 years of experience. While my employment history has taken many interesting turns over my lifetime, computers, and software specifically, has been something I’ve spent the most time doing professionally not counting my freelance and book writing projects.
I mention my background so you’ll understand that I’m writing here with quite a bit of professional background in the field even though I didn’t work much directly with AI technology itself.
When I posted The Atlantic article to my social media, a friend messaged me wanting to talk. He’s a young man in the early phase of his software technology career focusing mostly on technical writing and curriculum/training development. He was concerned.
He had asked to chat because he knows I'm informed about AI and he knew of my technical writing background. With the big story about ChatGPT churning all over the web, he was deeply troubled about the possible ramifications to his field.
I was honest. I told him I think he does need to be concerned about the future of his profession. I'm 68. Were I still actively in the field, I wouldn't be that worried. He's in his mid-20s. He should be concerned.
Of course, no one knows for sure where AI will take this field, or any other field or line of work, but here is what I believe will happen at some point, and probably sooner than later.
This is all punditry. I admit that. Still, I think my prognostications are at least within the ballpark of inevitability.
I believe within 5-10 years much of technical documentation will be written by AI. Certainly, the basic procedural stuff (Step 1, Step 2, and so on) will be written by AI, but even the contextual stuff surrounding the procedural documentation (use cases, examples, and implementation tips) will be written by AI eventually too.
It's not unlikely that AI will dive directly into the application code and extract the logic and write the corresponding documentation. Software code has a logic to it. AI is excellent at producing outputs that result from structured logic. AI is even good at crunching random data, but it’s really good when presented with a structured logic it can easily mine and add to its learning and output results.
Even if human technical writers are employed in that timeframe, I believe they will probably morph into editing and text enhancement writers that tweak the AI output with a more human touch and insight. Technical illustrators will add process flow and other graphic visuals. But I think even that's all temporary.
Do I think AI will also impact the creation of software and technical curriculum and training modules? Yes, I do. I think it will take longer for AI’s sophistication to deliver effective training to customers and users, but it will.
Also recently, the Lensa AI app made waves with its “magic avatars” feature that used AI to allow users to upload existing photos to create digital portraits of themselves presented in a range of styles such as “anime” or “fairy princess”. The results have been all over social media and you’ve likely seen some of them.
If such power resides in an app that costs a user a few bucks, imagine what power will reside in AI applications sold for big money to large corporations and eventually to smaller companies and businesses. There is big money to be made in eliminating jobs through AI. Where the money flows, history tends to follow. I’m sure the power to create curriculum modules, video tutorials, and other deliverables historically done by a technical communicator already exists, or soon will.
If you're a company and you know that you could eliminate an entire category of jobs and the burden of those salaries by utilizing technology, you'll do it. That’s always been the case. Automatic elevators replaced lift operators and the modern cell phone has eliminated the telephone operator. I'm not saying that such replacements of jobs is always a great thing. I'm not saying that it doesn’t do damage to many longstanding jobs and careers. But I think it's reality and it would be folly to think the emergence of AI will have any less an effect. It’s effect will probably be far greater and more pervasive into multiple fields.
Frankly, anyone in a technical field should be considering the implications of AI, and doing it now is wise so as not to be caught flatfooted and needing to make a sudden decision to change course.
Related to this is my consistent suggestion to anyone wanting to maintain a lucrative career going forward. Learn as much about technology as possible, even if you’re not working directly on developing or delivering that technology. Automation, AI, and robotics are going to change almost every profession and create many yet to be born. At the same time, embrace being a generalist and foster your self-education and learning skills because you’ll have to constantly reinvent yourself and the pace of such reinventions will escalate over time.
I hope this post isn’t bumming you out. Instead, I hope it serves to make you aware of what you need to do in order to remain gainfully employed as AI and other technologies run rip shod over entire employment sectors heretofore believed to be solidly secure. Forewarned is forearmed, as they say.
You can use this link to access all my writings and social media and ways to support my work.