ChatGPT has taken over the internet. So, why is everyone talking about the natural language processing tool? How should (or shouldn’t) we use it to automate our work? We asked two content designers to share their differing views on using artificial intelligence (AI) in the UX writing process.
Position: For ChatGPT
The following section was contributed by Patrick Stafford, COO of UX Content Collective.
I’ve seen enough supposed game-changing technology to know better than make a premature declaration about ChatGPT. But there’s a difference between “this will be big in 10 years” and something that’s producing useful stuff now, so it’s worth paying attention.
I don’t really blame the content design industry for being on defense. The message to entry-level folks has been something along the lines of, “love words? Can’t code? Become a UX writer.” I’m a proponent of that message, by the way. We’re all in this industry because we like writing and we appreciate language.
Our latest salary survey asks people what their favorite part of the job is. Unsurprisingly, one of the most popular answers is “solving problems with words.” It’s fun and satisfying to craft messaging under creative restraint.
But this defense against artificial intelligence, I believe, comes from a place of confusion about what our job is and does. Before discussing AI, we need to understand this: we’re designers first and writers second.
Reason #1: Content design involves more than string generating
Here’s the actual experience of content design: you spend 80% of your time looking at research, talking to people, attending critiques, understanding your users, looking at information, creating tests, and so on. By the time you get to the writing, you’ve already got guardrails on what the best user experience will be.
That means your writing process is just 10-20% of the job. Putting pen to paper (so to speak) is the very last thing you do, and it’s entirely informed by the process of design. There are creative elements along the way—you could say writing under constraint is the highest form of creativity—but you’re in the pursuit of organizational goals and not creative craft.
But still, we’re passionate about it. The passion we feel, though, is not through writing 50 ideas on Post-it notes and sticking them on a wall. The euphoria of creativity comes from connecting two previously unlinked ideas and discovering something new. It’s moving dirt around until you find some flecks of gold underneath.
Writing app or product copy is a lot like moving that dirt around. All ChatGPT is doing, is helping you move through that dirt a little faster. It’s a shovel.
Reason #2: Prompt engineering is creative in and of itself
The creative process is different for every content designer, and I can only speak personally. Whenever I need to create strings or copy for some process, I’ll do a mind dump. It’s a blast of terrible and obvious ideas to get through the surface-level content and into something deeper. I had a manager who used to prompt us to write “50 ideas in five minutes” on Post-it notes to get this type of creativity going.
This type of work is what ChatGPT is good for. Helping you come up with a list of strings that can then generate you to think of more ideas on top of that. It’s a way for you to start writing and coming up with something good—like any writing prompt.
This suggestion is often met with derision from AI skeptics who say the output isn’t great. Part of the issue with underestimating ChatGPT now is that so much of the output depends on how willing you are to understand how it works and why it works. That is to say, it’s only as helpful as you yourself make it to be. That’s exactly why prompt engineering is becoming more popular as a practice, to train the bot into giving you helpful responses. This means understanding how ChatGPT can be used to provide strings after giving insight on tone, character limits, and so on.
Of course, it’s up to you to decide how and when this output is used in your design. Sounds pretty creative to me.
Reason #3: Using new tools will only make you stronger
A lot of writers put value in their knowledge of grammar, syntax, vocabulary, and other types of performative declarations. Those are great tools but they’re exactly that: tools. You can write amazing copy without knowing a rule of grammar and you can write terrible copy knowing everything in the book. None of it changes your ability to understand an entire design process end-to-end. That’s your job.
It seems strange to incorporate so many tools but then stop at ChatGPT. Do you use Excel? A user testing platform? What about analytics tools? If you’re willing to use those, ChatGPT should just be another one in your arsenal. It’ll be hard to compete when other content designers are able to work at a faster pace.
Here’s the way I see it: it’s dangerous to become romantically attached to the way in which you perform your job. I’ve seen it way too many times and people get left behind. They don’t want to learn Figma, or they don’t want to learn Miro, and they can’t be bothered putting something new in their toolbox. Meanwhile, those who can wield these tools with confidence end up with an advantage.
Don’t be so rigid that you put yourself into obscurity. In 5 years, a content designer with the ability to use an AI tool will be worth more than one who doesn’t touch AI for any reason. You don’t have to like it. But know how to use it—and when to wield it appropriately.
Position: Against ChatGPT
The following section was contributed by Lauren Salisbury Douglas, Freelance UX Writer.
This won’t address the many subjects swirling in my brain about ChatGPT or the other AI tools currently available—ranging from the astounding and admirable to the dark and sinister. It won’t pontificate on the potential positives and negatives this tech has unleashed upon our world. I can’t address the ethics right now, because I’m still considering it, and because there continues to be a plethora of articles on the web that frighten me.
This post is just about one writer, staring at a blinking chatbot cursor, making the decision to put the screen down.
I’m a UX writer and copywriter, and I’ve worked across industries like consumer goods, luxury furnishings, health & wellness, finance, and software. Eagerly, I tried ChatGPT. It didn’t take me long to decide that I won’t use it for work, because ultimately, I’m not that kind of writer.
There’s no doubt that it could help many writers as a brainstorming tool, but I will never copy and paste this stuff even as a draft to edit because I’m not a copy and paster—I’m a writer. Using a chatbot takes that all away.
Reason #1: AI can’t empathize with users like we can
I liken UX writing to acting. That doesn’t mean I’m performing or being fake; it means I’m employing empathy. When I’m working on a project, I don the company brand and user personas like second skins, and from there, I act it out—or literally, write it out. While that may sound dramatic, this is the mindset I employ for all non-personal writing. It’s a strategy that gets me quantifiable results. I used a similar tactic in my previous career developing fragrances for consumer goods. While I worked, I tried to embody the person I imagined sniffing the body wash or candle I was developing. What did they like? What else were they accustomed to smelling in their daily life? Who did they want to be? How did they want to feel while experiencing this scent? What did I (or the brand) need them to feel when inhaling this blend I knew would instantly spark memories and build neural connections?
In short, empathy is what empowers me to create products successfully for a wide variety of industries. It’s critical for good UX writing, marketing writing, and web development.
The bridge to access my empathy is completely absent when I use ChatGPT. Even if using it to brainstorm ideas —for CTA button ideas, for example—I still question whether the results are appropriate or good enough to go into a product. (I asked ChatGPT to give me some CTAs for a particular kind of app I’m working on. It did, some of which were usable but not groundbreaking, and others were 5 or more words long, so I didn’t use them.) I’m full of sparks when writing, but dormant as an unlit match when “writing” via ChatGPT.
Reason #2: Writing without passion is pointless
And that takes me to the next crucial element that remains untapped if I use ChatGPT for UX writing: the passion! I’m passionate about writing because of certain feelings in my mind, body, and soul, too—an alignment; a gentle buzzing; a feeling of forward motion; the excitement of a challenge accepted; flow. All the things that make earning a living enjoyable!
The empathy and passion that I relish are not for their own sake alone. They are my data. They are my rationale and why I believe in my own work. They are the framework of how my own work is architected. I can write, take feedback, provide justifications, and move parts around, when it’s a structure I built myself on a skeleton of empathy and passion. It’s how I know why I have written what I have. With a chatbot, there’s no deep sense of ownership over my work. That means I’m not a writer anymore.
I take pride in the work I do and the products I create. I have a strong sense of self and highly value integrity. It’s unlikely I could use components or articles written by ChatGPT without feeling, well—like I’m cheating.
From a business perspective, of course, one will argue that only results matter. If bots are outperforming humans as UX writers, it doesn’t matter whether a random person likes to write or not.
But I would argue that it does. I don’t see AI outperforming human writers yet. Furthermore, the more automated our world becomes, the more human beings will crave not only empathy, passion, connection, and individual style, but also products created with these elements, even more. Global trends on a macro level tend to develop as reactions to what came before them, like the predictable swing of a pendulum. On a micro level, the deeper into our lives AI infiltrates, the more valuable actual human beings and their own work will become. Because unless AI becomes sentient, it can never be truly empathetic. That human business of seeing and being seen, of touching and being touched, emotionally, is ultimately what is most valuable, and fast becoming an increasingly rare commodity.
For the pleasure of comparison, let’s explore this article topic as a prompt for ChatGPT. I entered this prompt only after writing this post myself, so as not to influence my own work. You be the judge!
Prompt: “Write an article about how a UX writer will not use ChatGPT for her work because it lacks passion.”
OpenAI’s chatbot response:
“As a UX writer, my job is to create content that not only informs users but also engages and excites them. One of the key elements of good content is passion—a sense of enthusiasm and energy that can make the difference between a user scrolling past your product and deciding to engage with it.
Unfortunately, as a machine learning model, ChatGPT lacks the ability to convey passion in its responses. While it may be able to provide informative and technically accurate content, it lacks the human touch that is essential for creating a truly engaging user experience.
To create effective UX content, I rely on my own creativity, intuition, and empathy to craft messages that resonate with users. I can adjust my tone and language based on the user’s needs and goals, tailoring my content to fit the unique characteristics of each audience.
In contrast, ChatGPT relies on pre-programmed responses that lack the nuance and context needed to truly connect with users. It cannot adapt to changes in the user’s mood, preferences, or priorities, which can lead to a frustrating and disconnected user experience.
Furthermore, as a UX writer, I must take into account the specific branding, marketing, and messaging goals of each company I work with. I must understand the voice and tone of the company and craft messages that align with the overall brand strategy. ChatGPT, on the other hand, is not equipped to understand the nuances of brand messaging and may inadvertently create content that is off-brand or inconsistent.
While ChatGPT can be a useful tool for certain tasks, such as generating quick responses or providing technical information, it cannot replace the creativity, intuition, and passion of a human UX writer. As long as the goal is to create content that truly engages and excites users, a UX writer will always be the better choice.”
Is ChatGPT going to kill writing? No. I’m sure it could be a useful tool for many writers. But ChatGPT kills writing for me, and that’s why I won’t use it…yet.