How to effectively get started using AI-based tools as a designer
Acting on Jakob Nielsen’s stunning prediction of AI being the future of UX
Last week, I learned that Jakob Nielsen, the guru of Web Page Usability and co-founder of the Nielsen Norman group, talked about how AI is the future of UX.
Now that ChatGPT has been around for almost a year, I expected many larger organizations to have an opinion. However, I wasn’t expecting how supportive he was: he predicted a shortage of UX Professionals with two years' of AI usability experience in 2025, with each professional receiving 500 job offers.
At the same time, there's quite a lot of noise around learning AI-based tools. Nestled between get-rich-quick schemes and academic theses on AI are very few resources for non-technical people. As a result, many Junior Designers may not know how to learn this skill.
However, based on my experience in Data-Informed Design and AI, the answer is more straightforward than you think. To learn AI-driven tools, you must first adopt a new mental model centered around four words: Garbage In, Garbage Out.
Having the right mental model for AI is crucial
I've spent most of my UX career working in fields where data is king. From Healthcare and Federal UX to B2B/SaaS products, I've had to adapt design practices to incorporate Analytics, Metrics, and even AI-informed insights into my process.
This is why I'll tell you that the first place that Designers should start is with the mental model they should use around AI and data. To do that, I'll provide you with four words that have governed Computer Science (and related fields) for almost 80 years: Garbage In, Garbage Out.
In other words, if the quality of your input, data, or AI prompt is garbage, your output will be as well.
I first came across this phrase when I started learning Data Visualization, where it was of crucial importance. After all, it didn't matter if I designed the world's best data visualization: if that visualization was based on garbage data, it was a garbage visualization.
In addition, my garbage visualization may trick and mislead users into drawing incorrect conclusions.
Design tends not to be so black and white, so we're often unfamiliar with this thinking. However, it's crucial to understand this mindset because no fancy algorithm can spit out a perfect answer to a garbage question.
This is why the mental model most novice users (and Designers) take to AI-based tools can be so damaging. Why? They rely on the first mental model that comes to mind whenever they see a blank text field: they treat it like a search engine (i.e., Google.com).
You'll get a basic answer when you treat ChatGPT like a Google search and ask a basic question. How many of you did just that and thought AI was just a novelty?
However, there are better mental models to use. Instead, here's another mental model to consider: using AI tools as a sketch or design critique.
Generative AI tools 'replace' sketching or serve as a design critique
I'd be showing my age as a designer if I used the mental model of sketching for AI-based tools. While that's true, many newer designers don't sketch that much (or at all).
So, I'll instead use the mental model of the design critique. Imagine you're the reviewer and will be reviewing someone's draft that isn't very good at design.
This first draft, either writing or images, can give you a powerful headstart to your design process, but it must also be approached carefully.
One such reason is that these tools might cause you to miss the point of using them: AI-generated tools are for generating creativity and coming up with ideas.
For example, asking ChatGPT is an excellent place to start if you need some copy for your design. It won't be as good as if you asked a copywriter, but that's not the point: getting something down on paper to discuss with your team is the point.
If the copy is terrible, that's at least some feedback from your team compared to nothing. Likewise, MidJourneyAI can generate mediocre designs compared to what you'd create as a designer. However, it's a way to generate creativity for free.
After all, you may visit sites like Behance, Dribbble, or other places to inspire your design creativity. Using an AI-generated tool should be no different. Research around what man and machine are best at has been around for over 70 years, but it's only recently that this has shifted slightly.
While the verdict is still out, AI tends to be more creative than humans, while humans are better at novelty.
Let AI help you create better designs and augment your design process. After all, other skills will become a lot more critical in the future.
Generative AI will make thinking about design 10x more important
I will do something to showcase the importance of strategic thinking for designers.
First, I'll go to Designercize and generate a random whiteboard prompt as if I'm practicing for a whiteboard exercise on a design interview.
Then, I'll plug that prompt into Midjourney AI, generating several sample images.
After that, I should have aced that whiteboard exercise, right? Not quite. There are design issues with each of these images, but there's one fundamental question you won't be able to answer if you do this: "Why did you make the design choices that you did?"
If you haven’t made any design choices (and relied solely on AI), you won’t be able to answer that. This is why AI-generated images can work for sketching or design inspiration but won't work for your design process.
Understanding the Why is a crucial part of being a designer on both ends. When you talk with users, you don't always take their suggestions at face value: you need to understand the motivations behind their suggestions (i.e., the Why).
Likewise, understanding Why a business is motivated to add specific features when users aren't requesting them can help you figure out alternative solutions that meet both business and user needs.
You need this skill as a Designer to make it in the Design field. But with the popularity of AI-generated tools that can create beautiful (if nonsensical) images, it will be more critical than ever.
AI tools (like ChatGPT) are a UX revolution, not an AI revolution
While I've always been interested in AI, a particular moment sparked my interest in learning these AI-based tools.
Cassie Kozyrkov, the Chief Decision Scientist at Google, stated that the recent popularity of AI-based tools wasn't an AI revolution but a UX revolution.
It's true: many underlying AI technologies have existed for a while. If ChatGPT runs on the GPT-4 model, there were also GPT 1–3. However, it was only recently that effort (and resources) were invested into the UX of these tools and designing for usefulness.
However, designers shouldn't just learn AI-based tools to help their design process and use it to help themselves. Jacob Nielsen predicted that a million companies will need UX professionals with AI usability experience by 2025, and we're on track to have perhaps 2,000 experts so far.
This means that learning AI-based tools and understanding the user experience of AI-based tools can put you in high demand as a Designer. If you've ever been interested in learning about AI, now is a great time to start to advance your career.
You need to start by changing how you think about AI-based tools.
Kai Wong is a Senior Product Designer and Data and Design newsletter writer. His book, Data-Informed UX Design, provides 21 small changes you can make to your design process to leverage the power of data and design.
I really enjoyed this article, and it's made me look at generative AI in new ways. As something that should be in my toolbox as a good data analyst, but not as a replacement for my skills. I've used ChatGPT a lot for coding, but it doesn't replace me actually thinking through what I want to create/build.