Embracing professional redefinition
- The obsolescence regime
- An alternative path
- Challenges with professional redefinition
- Letting go of attachments to “being a writer”
- Expanding our identities
- 10 use cases for AI in tech writer roles
- Letting go of authorship
- Pushing my limits
- Upskilling in AI
- Thinking more radically
The obsolescence regime
First, a quick recap from my last post. In AI and APIs: What works, what doesn’t, I talked about the “obsolescence regime.” The gist is that competitive necessity will drive increased AI reliance, resulting in three phases: (1) First, writers who embrace AI gain a competitive advantage while AI-resistant peers fall behind. (2) Next, over-reliance on AI leads to skill atrophy, making writers unable to produce content without AI assistance. (3) Finally, AI systems no longer require human oversight and deem writers obsolete. Writers are discarded.
An alternative path
I argued that this dystopian outcome doesn’t have to be inevitable. Rather than resist AI or become entirely dependent on it, technical writers can redefine their roles. As James Bessen explains, new technologies often morph and reshape job functions rather than eliminate them outright.
For example, instead of just competing with writing documentation, writers can reskill to become AI documentation engineers. They can learn prompting, fine-tuning, and collaboration techniques to productively partner with generative models. Skills like prompt engineering, few-shot learning, and custom LLM training can allow writers to optimize and direct AI capabilities.
This point about redefining and reshaping our professional roles deserves more reflection, though, because changing one’s professional role and identity is psychologically difficult. How to embrace professional redefinition is what this post is all about.
Challenges with professional redefinition
Redefining your professional identity can be like discarding your old body for a new one without knowing beforehand what kind of body you’ll soon be getting. As tech writers who may have embraced technology from humanities disciplines, struggling to come up to speed technically, we can be stubborn about adopting new tools or workflows. We might prefer instead to build on those skills we’ve already learned.
Pivoting from hands-on writing to AI oversight poses an identity crisis, even an existential threat. If I’m no longer the one crafting each sentence, what am I? What’s my role? It’s far simpler to reject or dismiss technologies that disrupt our sense of identity. This defensiveness is partly why tech writer reactions to AI are all over the map, from dismissive and facetious to cautious and measured to over-embracing and hyper-optimistic.
Letting go of attachments to “being a writer”
Let’s assume that to truly escape the obsolescence regime, we have to reinvent ourselves professionally. How do we do it? For technical writers, the most radical change AI presents is shifting us away from writing expertise as our core identity.
From my sophomore year in college until now (nearly 30 years) my identity has been associated with writing. During my sophomore year, I enrolled in a Writing Fellows program that involved working as dedicated writing tutors for two classes a semester. As part of the Writing Fellows program, we had to write our literary biography, describing our origin stories as writers. How did you know you were a writer? That kind of essay.
During the program, this idea of “being a writer” stuck with me. I majored in English, got an MFA in nonfiction writing, taught composition at the college level, worked as a marketing copywriter, and then as a technical writer. “Writer” has been a constant thread throughout my career.
Now as AI tools emerge, they possess the ability to write—intelligently, coherently, and effortlessly. Do I still cling to my identity as a writer, even though it seems pretty clear that writing jobs (or portions of those writing jobs) will soon be replaced by AI technology? This is what I mean by the radical project that is self-redefinition. It’s not like changing a board game that we’re playing, from Monopoly to Chutes and Ladders. It might involve changing our entire sense of who we are.
Expanding our identities
It could be that the whole AI wave is just a fad that will fizzle into party trick memes and nothing else. Or AI could usher in the dystopian cyborg future that science fiction movies have been depicting for decades. If the former happens, professional reinvention will only add new skills to our existing ones. If the latter happens, professional reinvention will become a requirement to stay relevant.
Currently, it seems everyone is trying to figure out how to use AI in productive ways. I don’t have all the answers to that question. Currently, most AI tools are heavily restricted in the enterprise. However, through experimentation with the tools available and with content on my blog and API course, I’ve come up with a list of 10 tasks that you can probably do with AI tools.
10 use cases for AI in tech writer roles
The following are 10 tasks that I think have strong applicability with AI tools and documentation. For each of these tasks, consider not just the task you’re performing but the identity transformation that might be taking place on the inside. That transformation of identity is the more important goal.
|Develop build and publishing scripts||You are a developer|
|Understand the meaning of code||You are a polyglot|
|Format content (HTML, XML, YAML)||You are a web designer|
|Distill needed updates from bug threads||You are a partner engineer|
|Summarize long content||You are a voracious reader|
|Synthesize insights from granular data||You are a data analyst|
|Seek advice on grammar and style||You are a linguist|
|Arrange content into information type patterns||You are an information architect|
|Compare API responses to identify discrepancies||You are a debugger|
|Create glossary definitions||You are a lexicographer|
The transformation of our role depends on our proficiency with each task. And that proficiency can be magnified and amplified through AI.
My point with the identity column is that we can’t timidly cling to “writer” as our only label. Our new challenges demand fundamental identity shifts, with the goal of progressing from limited notions toward expanded possibilities. Each new endeavor can prompt a radical self-redefinition, depending on how far you want to take it.
Letting go of authorship
The core redefinition prompted by AI seems to be to distance ourselves from writing. I’ve struggled to let go of authorship, and to be honest, I’m still not sure that I should. It’s hard for me to let an AI do the writing instead. My early attempts using AI to write blog content produced poor quality, but I’ve learned techniques to blend AI content more smoothly.
With practice, I can probably get much better at prompting and fine-tuning, performing more of a director role than actor role. Should I be doing this? This change scares me. Not only for the possibility of having my writing skills atrophy, but even for optics. Will people see me as a cheat or plagiarist if I use AI-generated text? Will I still feel the same about writing that “I” have produced? What will the content look like that is half AI-written, half Tom-written? The scenario is an identity crisis.
But this is perhaps the very attachment I need to give up to evolve. My sense of self can’t be based on writing alone, given how easily AI tools can write. GPT 4 might sound wordy and Wikipedia-like now, but what about GPT 9 or 10? At some point, most human-written writing might be so inferior that people will no longer care who wrote it, they’ll just want the good stuff.
Knowing this, I feel I should reshape myself around new skills like AI prompting and optimization. I need courage to let go of old notions of my purpose. My identity has to be fluid, not fixed. But this is easier said than done. I’ve been building and refining my identity for decades now. And the current output from generative AI is still mediocre, making the whole endeavor potentially a flop.
Pushing my limits
I’m attempting some experiments at work to grow new skills. Some weeks I dive so deeply into build scripts and publishing automation that the title “technical writer” seems wrong. During those weeks, I’m doing full document engineering. With AI help, I’m writing code I don’t always understand. This has led to some tricky situations. When an engineer vaguely suggests enhancing the code, I might struggle to implement their recommendation. Or when a bug pops up, I might barely notice it and have trouble fixing it. But the scripts work, and they save me a ton of time.
Operating complex scripts I only partly understand—in order to automate docs—feels risky. I’m caught between the tedium of manual work and the uncertainty of handing control to partially understood tools.
Upskilling in AI
I’m also trying to get more familiar with Notebook LM at work. I’m testing its limits and working to better grasp how it functions. My goal is to get good at optimizing and directing Notebook LM when solving doc bugs. Recently for a bug fix, I fed Notebook LM 30 pages of docs and bug thread text, then asked it to generate the result I needed: a short summary of the fix for release notes.
Notebook LM produced exactly the update required, with barely any editing from me. Engineering approved it no problem. This felt odd—I wasn’t fully confident in the AI’s accuracy, but relied on engineering to validate it. They did review and approve it. Is this the future? Is this a good practice? What about my subject matter expertise? Was I trading long-term productivity for short-term gains? Then again, the needed text was only a few sentences, not an entire topic, so it’s hard to judge the outcome.
Counting on AI to distill key information made me uneasy. But as I evolve my role, discomfort is expected. Building skills in emerging areas like responsible AI use, even without a clear path, is crucial to avoid obsolescence. The way forward remains hazy, but we have to keep pushing through uncertainty. Or are we running towards a cliff? In abdicating the content generation to AI, are we numbing the articulate language skills we’ve honed over decades of experience?
Thinking more radically
To truly future-proof myself, I need to think more radically. Activities like training custom LLMs or developing chatbots could redefine my role but are hard to explore while doing my current documentation duties, as these activities are time intensive and immersive. And who doesn’t have an endless backlog of documentation to-do’s already?
Despite the lack of time, I have to remind myself: The tech writer role itself might have an expiration date, so we have to reinvent despite the difficulty—before that expiration date arrives. Before that date arrives, we have to figure out a new plan. But what? We can’t be caught off guard.
When I spend a few days trying something new that flops (for example, trying to use AI for planning and prioritization), I have nothing to show for it and fall behind in my regular work. But on the flip side, imagine the wins when it works!
To avoid obsolescence, dabbling in new skills won’t cut it. We need to dedicate time to redefining our role through high-risk, high-reward experiments. But what the experiments should be, exactly, remains unclear. At the same time, we can’t totally ignore our current doc work. We’re shakily straddling at least two worlds—an unsure present and unclear future. This is the position we all find ourselves in.
This essay wouldn’t be complete without acknowledging some AI assistance. For fun, here is the Claude thread that shows how I used AI to help with this post. For the initial draft, I tried to steer Claude paragraph by paragraph through the ideas I wanted to express. I’m not sure it saved me much time, though, as I ended up rewriting most everything. Interestingly, as the essay progresses, I seem to try to take back control by injecting increasingly personal anecdotes and a perspective expressing uncertainty and self-doubt. This may have been me pushing back against the machine. Claude also let me go in directions that I later regretted.
For a podcast on a similar theme, see [Podcast] AI and APIs: What works, what doesn’t.
About Tom Johnson
I'm an API technical writer based in the Seattle area. On this blog, I write about topics related to technical writing and communication — such as software documentation, API documentation, AI, information architecture, content strategy, writing processes, plain language, tech comm careers, and more. Check out my API documentation course if you're looking for more info about documenting APIs. Or see my posts on AI and AI course section for more on the latest in AI and tech comm.
If you're a technical writer and want to keep on top of the latest trends in the tech comm, be sure to subscribe to email updates below. You can also learn more about me or contact me. Finally, note that the opinions I express on my blog are my own points of view, not that of my employer.