UX/UI design in the rise of AI

AI tools open up a landscape of change and bring many opportunities to design teams. The importance of continuing to focus on iterations and quality checks to deliver user-centered products.

It wasn’t long ago that AI arrived on a mass scale, and it hasn’t stopped generating debate, especially in sectors that see these tools as a threat that could make some roles obsolete. Since ChatGPT and Midjourney became popular, various tech giants have joined the race for AI innovation, and professionals in every field keep being surprised by the increasingly complex tasks these tools can solve in a matter of seconds, automating processes that used to depend on specialists.

But there’s much more on the horizon, and it’s still early to gauge the true impact of this technology without resorting to guesswork. What can’t be denied is that, with new technological change, roles inevitably take on new meaning.

To be clear from the start: AI isn’t going to replace designers, but it will certainly define which skills are valuable in the future.

AI as we know it today is still successfully taking its first steps, and we shouldn’t forget that new technologies have limitations and room for improvement, which evolve with use and user feedback. Rather than focusing on whether AI can replace designers, we should identify AI’s potential to be integrated into design processes, and which tasks design teams can automate by applying it.

What we know from data

We don’t yet have the data to know for sure what effect AI will have in replacing some professions, and that discussion is unlikely to have a real impact on market decisions. What we do know is that generative AI improves employee productivity by 66%, averaged across 3 studies. The most complex tasks are the ones that benefit most from AI tools, and teams can find their greatest impact in activities such as generating content, gathering information, user data for research or first drafts when designing.

A big caveat to this figure is that it only applies to the tasks AI can support today, so in areas such as UX design, where there are still processes AI hasn’t covered, the benefit may be smaller. But the future sounds so promising that productivity on AI-assisted design tasks could rise by as much as 100%.

Of course, quantity isn’t a relevant figure without quality. In that respect, the results of the studies cited above show quality indicators on a par for tasks done with and without AI. One interesting finding is that using AI narrows the gap between designers with more and less skill and experience. That becomes relevant when we think about which skills will end up making the difference when hiring a standout designer.

Although these are early data, the first conclusions already herald a big impact from this technology.

Current AI use cases

As we said earlier, AI isn’t going to replace designers, but the designers who know how to use AI correctly (and ethically) will probably be the ones who gain the advantage, since they’ll have a set of tools that gives them the flexibility to adapt quickly to any change. That’s why it’s important to know what uses already exist for AI tools and how we can integrate AI into design processes.

AI has the potential to improve user experiences through automation and hyper-personalization. What current contributions give us a hint of that? Let’s look at some examples of how AI is affecting UX design:

• Content generation: AI can help designers create text, images, video or audio automatically or semi-automatically, from data or instructions. This can save time and resources, and make it easier to adapt to different languages, formats or audiences. Some tools that use AI to generate content are GPT-3 and Lobe.
• Interface optimization: AI can analyze users’ behavior and preferences to propose improvements to interface design, such as color, typography, layout or navigation. This can increase usability, accessibility and conversion, and allow more efficient A/B testing. Some tools that use AI to optimize interfaces are Adobe Sensei, Google Optimize and Optimizely.
• Personalized experiences: AI can use users’ data and context to offer more relevant, satisfying experiences, adapting content, features or recommendations to their needs, tastes or emotions. This can improve user retention, trust and loyalty, and create added value. Some of the giants that personalize experiences are Netflix, Spotify and Amazon.
• Building virtual assistants: AI can create systems that interact with users by voice or text, providing information, assistance or entertainment. This can improve the accessibility, efficiency and empathy of services, and create new ways of communicating and building relationships with users.

A new paradigm in UI

AI’s impact won’t only change UX/UI design processes; it will also transform the way we interact with interfaces in general. At least, that’s what Jakob Nielsen says: he sees generative AI as a new UI paradigm, the first in 60 years.

Nielsen points out that there have been three distinct paradigms in user interfaces since the birth of computing. The first is batch processing, which appeared around 1945, when there was no back-and-forth between user and computer, just batches of information sent on punched cards to data centers, which returned printed sheets in response to the request.

The second appeared around 1964: command-based interaction. Here the user issues a command and takes turns with the computer, which executes another in response. This paradigm has dominated the last 60 years, evolving through the graphical interfaces that became widespread from 1984 with the Macintosh.

But the third paradigm, the one apparently being born with generative AI, is intent-based outcome specification. With one caveat: Nielsen doesn’t believe today’s tools (ChatGPT, Bard, etc.) represent the UI we’ll use in the future, since they still have major usability problems.

In these systems, the user doesn’t tell the computer what to do; they tell it what result they want. In these “prompts,” the user doesn’t specify how to achieve the result, reversing the locus of control of command-based interfaces. The problem is that the lack of information about the process, the lack of transparency and the biases of today’s AI make it hard to identify and correct errors in the results.

For Nielsen, it isn’t clear whether these generative AI systems can achieve high levels of usability, so they may end up coexisting with command-based systems, or mutating into hybrid interfaces.

Regardless of that still-blurry future, we can’t deny that this new way of interacting with systems that AI brings raises new ways of approaching technology, and will undoubtedly create new dynamics in processes as we know them today.

The human factor

The big obstacle for today’s AI is that, for us to trust its systems as a replacement for people, these tools need to show human values, ethical regulation, transparency, privacy controls, information security, traceability, fairness and inclusion, context and countless other characteristics that define the human experience.

In UX design, informed decisions require experience, common sense and knowledge of real-life contexts in order to connect, across iterations, with what users want from our products. There, AI can’t yet compete with humans.

To create truly intuitive interfaces, human vision will remain fundamental. Most likely, the bulk of design teams’ work will shift from the first iteration to the second or third, since processes will speed up with the arrival of generative AI that can create first sketches with more quality and in less time than before.

But if there’s one constant in the field of UX design, it’s that there will always be a need to iterate, test, gather feedback and more, to understand what works and what doesn’t. In that sense, generative AI is good at precisely that, generating ideas; but it can’t refine the product to take it further.

That’s why, although designers’ visual skills will remain important, the value organizations place on different skills will most likely change.

If you’d like to continue the conversation about AI and UX design, we recommend listening to episode 23 of our Digital Experience by Soho podcast: “How AI is revolutionizing user experience design” (in Spanish), available on all platforms.