AI Writing Tools: 4 Human-Centric Approaches
How managers and writers alike can use AI writing tools in a human-centric way
If you’re a manager, you’ve likely been trying to get a sense of how AI writing tools like ChatGPT will change how you hire content writers, and how you gauge their output. If you’re a writer, well, you’ve probably been biting your nails to the quick.
The truth is, AI writing tools will have an impact on the industry, just not in the ways you might fear. In many respects, they can be used to augment human-centric writing processes and create a more efficient workflow for both writers and managers alike.
It’s easy to understand why so many people are concerned that this development means the end of freelance copywriters and others who’ve been paid for the pen. But on closer examination, that’s not the way the current shift is unfolding: the big threat isn’t that AI writing tools will replace writers; the coming wave is that humans who use AI writing tools well are going to out-produce, on average, those who don’t know how to use them.
So, how can managers and writers alike take advantage of AI writing tools while still preserving the unique human touch? First, it’s important to understand the capabilities — and quirks — of AI writing tools.
What are AI Writing Tools and What is Different About Them?
I’ve been using an AI writing tool for a couple of years now. When you first start to use it, it feels like you’ve just upgraded from trimming your lawn with scissors to a riding mower. The productivity boost feels enormous but that power can quickly take over as if it had a mind of its own. If you want to avoid running over the flower garden, you have to be in control. It’s dangerously easy to use it badly, and the results can be a real mess.
ChatGPT or any of the other generative apps are high-performance tools, and they offer fantastic productivity gains — not unlike the productivity gains that came from the word processor — and using them is a skill that takes getting used to. When you just start out, it can feel like the AI tool is really doing the heavy lifting in the writing work. But when you look closer you realize that to get good results, you still need to be involved — and if you want your output to sound like you’re, y’know, a writer, you still need writing skills.
If you use the tools in a thoughtful way and stay present in the process, AI can nudge you along and help you express what you are trying to say. And yes, if we use our human traits of consciousness and intentionality as we use the tools, both the process and the results can actually be quite human-centric.
How to use AI writing tools: managing the process
When it comes to any kind of automation, managers get intrigued with the idea of cost savings. That’s understandable, but you have to keep in mind that the tools don’t do it all. Writing for humans is still a human endeavor, and writing using AI writing tools, while generally faster and more productive, requires effort and an emerging skill set.
It’s also important to create a system for evaluating AI-generated content. Just because an AI tool produces text doesn’t mean it’s perfect or even good; you still need to review and assess the output. Establishing clear criteria for evaluation will help ensure that everyone understands what constitutes acceptable. You’ll need to offer some kind of brief as input, as you would with any project, and provide constructive feedback accordingly.
Finally, it’s essential to have a plan for feedback and revision. AI writing tools can help writers get started quickly — but no machine-generated text is going to be perfect on the first pass. Working with an AI writing tool should be seen as a process of refining an idea rather than completing a project in one shot.
For managers, AI writing tools can be incredibly helpful for setting up templates and other guidelines to ensure consistency across all projects. However, they also need to remember that they are dealing with people. AI tools can help move the process along, but it’s important not to lose sight of the human touch. Offering feedback and having conversations about how the content will be used can be invaluable for encouraging experimentation and collaboration in order to produce better results overall.
Writing with AI tools: the guidelines to keep in mind
If you want the output and the process to be human-centric, then the human — that’s you — will have to be aware and in control while using it. It’s like water skiing (or at least: how I remember water skiing from when I was a kid). The boat is doing an awful lot of the work in pulling you forward, but you still have to be very aware of your balance, your surroundings, the tilt of your skis, and so on. (See: whether you appreciate that analogy or not, you can one hundred percent tell I wrote it myself. I think we can safely assume that AI writing tools don’t yet have any meaningful sense for what skiing feels like.)
REMEMBER TO DO THIS EVERY TIME:
- Review, review, review. Note that the common step in all of these approaches is a close and careful review.
- If there are statistics, names, or other “facts” in the generated text, check them. It’s very likely the “facts” are wrong in context, but they often make a good prompt for where citing a correct fact would be appropriate. Do your research and swap in the right information.
- If there are clever turns of phrase, search them to make sure they’re not already in use. Committing plagiarism by proxy through an AI tool is still plagiarism.
What I’m saying is: Do not uncritically accept what the AI tool writes! Much of the way these tools generate the output they do is by devouring massive amounts of readily-available content, such as blog posts on the open web (not unlike this one). Their learning process consists of algorithms that make sense of word relationships, so they’re able to recombine and put words and phrases together in original ways based on existing patterns of language use. But that doesn’t mean the tool might not string together a two- or three-word expression that may sound new to you, but which may already be in use in ways you aren’t aware of, or it might be the very phrase (like “tech humanist” or “human-centric digital transformation“) that someone who coined it has a search alert for. I can tell you that my alerts have been pinging constantly these past few months with posts which are clearly nonsense. That’s not the kind of experience anyone is looking for.
An AI tool might also inadvertently introduce biased language or implications into the writing, although humans are plenty capable of the same. Reviewing for inclusive language is a good step no matter how your content is created.
Four human-centric AI content writing work approaches
So what does that AI-augmented but human-centric writing process look like in practice? Well, I’ve noticed a few different approaches over these past few years. I’ve outlined them briefly below.
Draft mode –
- Write the clearest prompt you can. This is not all that dissimilar from writing a good creative brief, or asking a good strategic question. You have to do the work of articulating a starting point or the results will be generic and all over the place.
- Let the AI tool generate a response. Don’t let it go too far without review or you can get overwhelmed trying to organize the chaos.
- Review it all with a critical eye. Did any of the output get at what you meant to say? Did it bring you closer to a central point or farther away? Can you edit it all down to a core idea?
- Reword the prompt if needed. Keep trying to get closer to a strong central concept.
Note that sometimes seeing in print what you aren’t trying to say is an eye-opening step on the way to being able to express your thoughts more precisely.
Refine mode –
- Take a bit of text you’ve written (I regularly use Otter.ai to record voice notes and transcribe them as starting points for refinement) and ask AI to rewrite it to align with your objectives: clarity, simplicity, humor, style.
- Cut, edit, or move words around until the output reads like you intended it to.
- Try again and see what the AI produces.
- Make sure there is an appropriate balance between automation and human intervention.
This approach can help shift the focus from being a writer to being an editor, leaving you to finesse the AI’s output as you go.
Experiment mode –
- Take two or three different pieces of content and try mixing them up with different words
- Review the results on their own terms – does it make sense? Does it add something new to the conversation?
- If the AI tool generated something off-topic or unnecessary, delete or rephrase and try again.
Using the tool in this way can push you to come up with creative, offbeat content.
Ideate mode –
- Write a prompt for the AI tool to create a list of blog post ideas, tweet ideas, video titles, etc.
For example, “blog post ideas for leaders about the future of work” results in:
– Leadership in the Age of Automation
– The Challenges and Opportunities for Leaders in a Technology Driven World
– How AI is Transforming Leadership and Management Practices
– Keys to Being an Effective Leader in the Digital Age
– Creating Resilient Organizations with Intelligent Technologies
- Ditch any ideas that don’t work. (Get used to killing bad ideas fast.) Use the remaining list to prompt for more ideas.
- If any of the ideas are relevant and useful, try writing a brief outline or rough draft of a post using one of the methods above.
This approach can extend your brainstorming range, but make sure you emphasize the human aspects of content creation – editing, revising, and proofreading – as well as adding your own voice to what you write.
Getting used to the AI-augmented human work future
I’m not going to lie: this does takes practice. We all have a lot to learn. As a writer, using these ai writing tools well will require you to flex both ‘sides’ of your brain: creativity and analysis. As a manager, you’ll need to give clear briefs and hone your editorial skills. And while you’re at it, you’ll also want to use these tools yourself, so you’ll be acquiring a fresh set of skills to be able to use them in an effective and meaningful way.
But whatever your role on the team, by keeping the humans at at the center — both the humans you’re writing for and the humans steering the writing process — you can create content that stands out for its distinctiveness and insight.
The key is to keep a watchful eye on the process, check facts, make sure the tone stays true to what you want it to be. And just remember: at least for now, AI writing tools are only as good as the humans using them. So learn to use them well. Learn to manage humans who are using the tools. The productivity boost we get from these tools will create the next big wave in the future of work; we’re all better off learning to ski with style in the wake.