How junior designers can use ChatGPT without risking everyone’s trust
How to make use of the productivity boost AI offers without running into issues.
AI-based tools like ChatGPT have posed an unfair question to the UX community: Do you have enough experience to use it effectively without losing everyone’s trust?
That’s the view Jakob Nielsen, co-founder of the Nielsen Norman Group, has taken. He states that “AI Is Safest for Experienced UX Professionals” since you need enough experience to include a heavy dose of human judgment to incorporate AI into your Design workflow.
This is because UX fundamentally hinges on trust. Both the user and your team need to trust that your Design is based on best practices, evidence, and intuition instead of from an AI prompt.
So the unfair question AI poses to Junior Designers is simple: can we still trust your work if we find out you use tools like ChatGPT and MidJourneyAI within your work process?
To understand why that’s important, we need to look at UX’s history.
UX has been establishing trust for decades, and AI challenges that
UX and Design have always been future-facing, as we’re often designing features and products that have never existed. In addition, many of the products we create require a fundamental shift in user thinking.
This is why building trust has been a crucial part of the design process. Whether introducing users to inputting their credit card information online in the 90s or getting them to hop in strangers’ cars with Uber, we’ve always needed to help establish trust as a part of our design process.
Generative AI, like ChatGPT and Midjourney AI, can potentially ruin that trust. Not only users but businesses lose trust in websites, content, and more created entirely by AI.
Right now, it’s also one of the surest ways to ruin your reputation as a content creator.
However, the reality is that these tools offer such a massive boost to productivity (around 33%, according to one case study) that it’s hard to ignore the benefits.
This is where the unfair question comes from. Senior Designers often have enough judgment and intuition to be skeptical about AI to pick out the good parts of the responses while understanding where the AI tends to hallucinate or mix things up.
However, if you’re just getting started with AI and also a Junior to Mid-level Designer, does that mean that you can’t use it? No. That means you must reframe your thinking about the AI tools you’ll use by giving it the correct persona.
Imagine AI-based tools like a brilliant but disorganized assistant
The right way for Junior Designers to think about AI-based tools is to give them the persona of a brilliant but disorganized assistant.
You have probably seen this type of character through books, TV shows, and movies. You may even know some real-life people that are like this. They’re book-smart, awkward, and brilliant, but they’re also clumsy and disorganized.
Now imagine this person is your co-worker and assistant. They’ll help you work much faster, but you must supervise them. It’s not as though you shouldn’t trust their results: most of the time, they’ll probably be correct or in the general vicinity.
However, if you trust them completely, it will blow up in your face. Whether they create disorganized lists, have biases based on their training data, or create a completely inaccurate technical article, there are bad things that result from trusting them completely.
To give an example of how you might supervise them, we can turn to chatbots. While chatbots can provide a ton of feedback and help cut down on customer support response times, users have trust issues around how much we would trust AI-based chatbots with our details.
This is why Chatbots are usually for general help and inquiries, with them transferring you to a real agent if there are specific questions or details the user wants.
This is part of learning what you can trust AI to do and what you need to do yourself. Here’s how to get started.
How to approach AI-based tools as a Junior Designer
Understand ways that AI can help your design process
There are too many potential tasks that AI can help you list in an article, so I’ll give a quick overview.
AI-based tools help Designers in 3 main ways:
Writing/Planning Content
Working with User Research
Designing first drafts and sketches
Here are some examples of each category that you can act on.
Writing/Planning Content:
Planning a workshop agenda, a meeting, or something else
Proofreading and rewriting any content that has been produced
Working with user research:
Writing user interview questions or a research plan
Doing sentiment analysis to understand initial themes in the research
Rewriting or summarizing key points to make it more easily understood
Understanding the “Why” behind user actions and what to do about it
Designing first drafts and sketches:
Create illustrations for design artifacts like personas and journey maps
Generate ideas for creativity while also applying certain constraints (i.e., “Design in a Center constrained grid, etc.”)
Summarizing best practices to provide evidence for design recommendations (“i.e., proof that dark patterns hurt long-term retention”)
Understanding job roles and the story you need to convey to them.
I talk about more tips and tricks around AI and working with Data in my newsletter if you’re interested.
Take your time explaining the context.
UX professionals are infamous for always saying, “It Depends” in response to specific questions. This is because context matters heavily to the field of UX: a search bar with filters meant to filter treatment drugs for doctors may require a completely different design than one used to filter out food delivery options.
This is why establishing context within a prompt matters so much. If you need a ton of context, so does your virtual assistant working with you.
Fortunately, ChatGPT has rolled out custom instructions to help with this process. There may be a ton of context that you want to give, and some may be the same across every prompt you provide.
Custom instructions allow you to apply that particular context to every prompt you create, which helps avoid a lot of tedious copy-pasting. However, you’ll still likely need to provide some additional context.
For example, if you want to craft a research plan, you’ll need to provide context around things like the type of study, budget, resources, timeline, and more.
Treat AI as a starting point.
One of my strongest beliefs for all UX professionals at all skill levels is that AI is a great starting point. A first draft, a basic sketch, or some default copy can be a great way of jumpstarting your design process. Likewise, a summary of user research insights can give you some potential themes.
However, it can’t be the completed product for the reasons I’ve listed above. You can’t trust AI with the finished results unless you want to end up in hot water and risk your users and businesses’ trust.
Instead, think of it like getting a jump start on your design process: having a highly polished 1st draft for you to review, edit, and make changes can allow you to work much more efficiently.
AI helps to close the skills (if not experience) gap
Let’s face it: UX is in a weak job market right now, and from what I’ve heard from mentoring aspiring designers, it can feel challenging to break into the field.
However, learning about AI can help you figure out how to bridge the skills gap. A study done by the Boston Consulting Group showed that AI benefited those with the least skills the most (the 33% boost that I talked about earlier).
While this is only a small sample size, studies seem promising that AI can provide a massive boost to your skills and productivity if you know how to use it.
Spending time addressing the unfair question and learning how to approach it can result in you having a leg up in a rough job market.
So, if you’re scratching your head trying to figure out how to apply UX to your field, consider this approach. It might just provide you with the information you need.
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.