7 posts tagged with "AI"

The AI revolution is laying bare so many biases. LLM bad behavior and ChatGPT hallucinations revealing the chat room cruelties in its learning data. Research and lawsuits keep piling up around how AI job screening tools discriminate illegally, be it age, gender or ethnicity.

The biases against technical writing are a little bit more complex. For so many decades, the hallmark of great technical writing was that it was invisible. It seamlessly supported people and the author was not obvious. Now these humble habits of best practice mean we're not in the training data. The corpus of training data LLMs rely on to understand speech (regardless of what the question is) skews heavily towards sales and marketing brochures and does not see the invisible effort of so many content creators. The tip of the iceberg is mistaken for the whole thing.

The release of the State of the Docs report 2026 is a snapshot of where the industry is and provides a vital path forward to a better docs ecosystem in the coming years. Over a thousand people responded from across the world of documentation. It's such a timely and important document as our content landscape radically shifts. It's an honor to be included as one of the people interviewed and quoted.

Back in 2018 I gave a presentation at Content Connections, titled Governance in Content: Draconian Overlord or Supportive Coach. This was back in an era when most folks didn't really think about AI.

Part of the question I wrestled with was, what is AI? What is intelligence? I went with John McCarthy's definition, "Intelligence is the computational ability to achieve goals in the world." Technically this means that a thermostat has intelligence, but its intelligence is somewhat limited.

The heart of my argument, however, is that AI had the ability to take care of boring, repetitive, "CLEAN YOUR ROOM!" mum stuff for content creators. Through this collaboration, technical writers, editors, managers could spend more time coaching and working with people to make great content.

For some generative AI has felt like a liberation. At long last they can communicate in a way that is heard and understood.

But it's useful sometimes to pause and think, liberated from what? When something is used as an assistive device it is handy to think why it is needed? What other factors are in play?

Generative AI is a statistical aggregate machine. Sometimes it can be good at surfacing what should be good practices that too often aren’t followed (like when a manager makes a dodgy decision by themselves, when talking to any HR person might help them follow best practice). While it can be useful, AI is best at continuing the way things have been done and smoothly perpetuating dominant hegemony.

When someone might feel the need to mask or code switch their language using AI, I think it is up to us to self interrogate. Why might this person want to smooth down their individual edges?

IQ predicts how well you'll do in your school years and how much salary you can make over the course of a career because it says what job you can get... but once you're in those professions everybody else is about as smart as you are, and that is where emotional intelligence kicks in. - Daniel Goleman

Many technical writers are fairly empathic people. To be good at our job, we need to care about user experiences, we need to be able to imagine different scenarios and meet different needs. I know that I am sometimes better at understanding the needs of others more than my own. Other peoples needs are interesting puzzles, my own needs can feel more like irritants that get in the way of me doing stuff!

But, it turns out, if you don't honor all the aspects of emotional intelligence you're selling yourself short and creating problems for yourself. As Ramona Hacker points out, too much empathy can sometimes be a barrier to full emotional intelligence. The noise of so much empathy can drown out too many other parts or give the illusion of strong emotional intelligence when there are aspects that need more work.

If you want to jump straight into the Agenda and Minutes Template:

• Create your own copy
• Master document, where you can leave comments and provide feedback

Minutes are important, and not just how we choose to use the finite amount of time we have in life. Minutes are recordings of decisions and discussions that occurred during a meeting. Done right, they are powerful and meaningful. As a technical writer, I've often used an agenda and minutes to drive decision making, make next steps clear and to manage my own enthusiastic use of workplace chat!

I have two strong opinions (aka hot takes? but is it really hot?). One is that technical writers have a vested interested in good meeting minutes and should just start writing them if the minutes are falling short.

If you want to jump straight into Search Term Analysis spreadsheet:

• Create your own copy
• Master document, where you can leave comments and provide feedback

About​

When we talk about content quality, findability is an important factor that extends beyond clean architecture. "Can the user find the data?" is a common question. Can the user find the data if they start with a Google search takes us deeper into this investigation.

At the Write the Docs Conference in Portland this year I gave a lightning talk on "Google Magic!" It showed how and get insight into your documentation in five minutes. Through this approach you can get a clearer understanding of your users and identify the most impactful ways to improve your docs.

This article goes into a bit more detail and provides a template for conducting several kinds of search term analysis.

Jump straight straight into the Neurodivergent Quality of Life Prioritization (NQLP) Matrix:

• Create your own copy of the NQLP
• Master document, where you can leave comments and provide feedback

Creating a framework to manage the brain weasels​

Whenever life doesn't go to plan, there is an opportunity to self reflect and find opportunities to learn and grow. This is a useful habit to cultivate and can turn adversity into an alchemical experience.

What is not helpful, is to look at the data, look at all the variables that go into being a human and come to the conclusion that you are a trash human/need to improve across all the criteria. This is especially challenging when you're not even sure what the criteria are, or perhaps the criteria changes as you analyze yourself from different angles.