While artificial intelligence (AI) as we think about it now has only existed in the public eye for a year or so—AI has actually been around far longer. From the first digital programmable computer in the 1940s, Alan Turing’s famous imitation game in 1950, to Apple’s Siri in 2016—machine learning and AI have existed for decades. The biggest changes arise in how we are using and interacting with AI.
With discourse everywhere around the future of UX research and product development in an AI-led era, we wanted to get to the heart of the matter: what does the future of UX research look like?
Collecting insights and predictions from experts at Notion, Meta, and Dovetail, read on for some future-gazing to shed light on this question.
The artificial intelligence glossary
Let’s start with some definitions. First up, AI: we all know this one—it’s ‘artificial intelligence’.
In plain English, the true definition of AI is Artificial General Intelligence (or ‘General AI’)—the simulation of human intelligence by a machine. This kind of AI can understand and learn as humans can, but to be created, it would require machines to become conscious, with full cognitive ability.
In a sentence, AI is just really good pattern recognition based on vast quantities of data.
Desiree Conceicao, Software Engineer at Dovetail
As Desiree Conceicao, Software Engineer at Dovetail, explains, “Real artificial intelligence, where a machine can learn and act independently—the way humans do—is still a distant dream. Instead, most of the discussion about AI at the moment uses ‘AI’ as a synonym for machine learning.”
So, what is Machine Learning?
You’ve almost definitely already been using Machine Learning (ML) day-to-day. It’s used for everything from predictive text to facial recognition, transcription, voice-to-text, and much more. In short, it’s the development of computer systems that can learn and adapt without explicit instruction, by using algorithms and statistical models to analyze data and infer patterns from it.
AI and ML are the two big players in AI terminology, but there’s some other need-to-know that might be useful:
- AI model: A program or algorithm that can recognize patterns and make predictions or decisions based on training data it has been fed.
- Generative AI: This is a type of AI that can produce new content, such as text, imagery, audio, and (synthetic) data. Advancements in generative AI are what’s largely responsible for the recent wave of interest, investment, and awareness of AI.
- Neural Network model: This AI model teaches computers to process data in the same way as a human brain, so it can make decisions on its own. (Unlike a Machine Learning model, where the computer makes decisions based on what it's learned from data.)
- API: This is an Application Programming Interface—a type of software interface that acts as an intermediary to allow different computer applications to talk to each other.
- NLP: Standing for Natural Language Processing, NLP is a type of AI that can conduct communication between humans and machines. Like other AI, it’s trained with real-life data and then uses pattern recognition to come up with appropriate responses. NLP is how chatbots work and is often used for translation and transcription.
- Foundation models: Created to do general tasks, ‘foundation model’ is a loose term given to models trained on vast amounts of data that can then be adapted to a variety of applications.
- LLM: Large Language Models use language as their input and output, have many parameters (functions), and have been given an immense quantity of data to train with.
- GPT: Generative Pre-trained Transformers is an umbrella term for a series of neural network models that use transformer architecture. Desiree explains that “GPT models marked the first time a model could be given commands in natural language and reply with human-like text.”
With the GPT3 model powering apps like ChatGPT, the GPT series is a key advancement in AI, making AI and machine learning far more accessible to the general public than ever before.
Here is where the recent emergence of AI excitement really begins.
So, now we’re on the same page about what AI is, let’s get into its potential applications, and how it's shaping the world of research and product.
For the sake of clarity, when we talk about ‘AI’ in this article, we’re broadly referring to generative AI and machine learning. After all, diving into every kind of artificial intelligence and machine learning, from Deep Learning to true AI, is a far bigger topic…
The opportunities of AI
From current uses in accessibility technology to developments in driverless cars or the North Star of AI-powered brain surgeons, the possible uses of AI are endless.
Just a handful of the potential applications of AI include:
What if AI could predict tornadoes or tsunamis? The premise of a perfectly-accurate or early-detection weather system doesn’t seem that far off when we consider the amount of training data that could be provided, and the advancements in weather prediction technology already. We could even see technology like AI-enabled traffic lights or SatNavs that collect data on local pollution and make real-time adjustments to traffic flow. Or what about idle monitoring of endangered species without human intervention, or automated renewable energy?
AI has massive implementation potential for healthcare, from using technology for early symptom detection or advanced diagnostics, to the creation of antidotes and new medicine production. Or what about AI-powered assistants for precise but effective brain surgery, or at-home caring needs for the elderly?
Protection and de-risking
Imagine if high-risk jobs like collecting nuclear waste, bomb disposal or fire-fighting could be conducted—and better yet, prevented—by technology. It may feel far off to have robots putting out fires, but the high-performance, high-accuracy, and high-speed abilities of AI make it prime for these high-risk and high-stress situations. Even self-driving vehicles, which are currently in development, depend on AI, and can help make travel more accessible and dramatically reduce road accidents.
From AI-generated closed captioning and audio descriptions, to text-to-speech functions on computers—accessibility is one element of technology that has already benefited hugely thanks to AI. But consider what further AI advancements for accessibility could look like: what about transforming in-the-moment conversations, content, or reality into audio or visuals? How about eye-tracking that can function as a computer mouse, AI that can generate quality, accessible coding, or can translate human speech into multiple other languages in real-time?
Whether you’re an AI advocate or a wary skeptic, it’s undeniable that there is vast potential in AI. What’s more, there’s a clear throughline in the purpose behind how AI tech comes into play and its potential applications. Ultimately, AI always comes back to a fundamental purpose of helping people accomplish tasks more efficiently and effectively.
In this sense, AI is the same as any other technology. While it may feel counterintuitive for an analytical and algorithm-based technology, at its core, AI is user-centered.
Zooming into the tech industry, Jo Widawski, CEO & Co-founder of Maze, feels the emergence of AI has sparked excitement among product leaders and researchers, becoming a kind of beacon amid the gloom of the current market and wider economic difficulties.
I think it (AI) kind of ignites and infuses a renewed sense of tech-led passion.
Where the recent recession has forced researchers to adapt, AI’s rise brings forth a strategic rethinking of the function of research and the role of researchers in our organizations. As Jo says, "Such a disruptive technology forces us to think strategically about the impact—how it affects the overall organization design and power dynamics."
AI has initiated these important conversations around priorities, and continues the motion of research becoming a strategic partner in organizations looking to future-proof their success.
AI is part of the solution, not the whole solution
While AI offers many exciting possibilities, it’s not an all-encompassing solution. Linus Lee, Research Engineer at Notion, emphasizes that generative AI has huge potential: “There’s a lot of positive things going for this kind of [language model] interface. For one, it’s very flexible and very adaptable. You can ask the AI for anything, and as long as the response is some text or image format, it can get back to you.”
Linus also comments on the flexibility and accessibility of LLM models, remarking that they are: “technically simple to implement, and the UI design is a fairly light layer on top. Chat interfaces like this are also quite easy to use, as most people know how to use a messaging app.”
Nonetheless, Linus remarks that “There is still a large interface gap with what LLMs provide and the kind of solutions that really fit into users’ workflows.”
At Notion, Linus and his team are focused on bridging this gap—aiming to “Close this gap between the computer and the human, and strike a balance between the intuitiveness of graphical user interfaces, and the flexibility of natural language paradigms.”
In short, the current AI tools are ideal for teams looking to blend human expertise with AI efficiency. However, we're a long way off true artificial intelligence, where AI could ever replace humans. Which is exactly why research and UX will simply always be needed to support AI's creation and iteration—because, fundamentally, it is about understanding people and solving their problems. While AI may be the conduit, user experience is the pathway.
The importance of UX in the age of AI
Whether you’re experimenting with chatbot-generated research questions, or using AI to analyze moderated interviews, an undeniable benefit of AI is its ability to process data quickly and expedite manual workloads.
These developments mean AI is freeing up time and space for product teams to focus solely on the user, rather than the technology itself.
While UX has always been about how the user and product interact, the advent of AI means UX is evolving from interface and interaction design into a fully developed two-way conversation.
This two-way interaction demands a more profound understanding of user intent and context than ever before—making discovery and research central, and mission-critical, to the practice.
The role of UX has always been sitting between product discovery and delivery, with practitioners generally having limited bandwidth for the former, explains Jo: “I believe AI is going to dramatically change that ratio: we’re already seeing the delivery process becoming AI-assisted, and we can anticipate that this will redefine the UX role and allow practitioners to focus heavily on discovery and research to identify and solve user needs.”
With generative AI being widely available, Jo comments that the way organizations and products can differentiate themselves comes down to the user experience. After all, if anyone can build with AI, what makes your product worth using over another?
Time-to-right is not about going to market fast, but about how fast your company can identify and solve user needs—making user insights the new center of gravity for the successful organizations of tomorrow.
Here is where AI tools enter the product team's tech stack: using AI tools to scale and expedite learnings.
The new role of research
Rather than being a threat to UX practitioners, AI can be a powerful tool to augment their capabilities and efficiency. But how do you get started? And what can teams do now to prepare for AI and get ahead of the curve?
Let AI do the heavy lifting
According to Benjamin Humphrey, CEO and Co-founder of Dovetail, the introduction of AI into research workflows is about considering “What you want to hold onto and what you want to get rid of”. Benjamin notes that “Most of the excitement today is about how [AI] can replace the analysis and synthesis process.”
As AI tools take over the manual labor of research—be that data management, reviewing questions for bias and grammatical errors, interview transcription, or automating report sharing, this alleviates the workload for research and UX professionals.
In terms of preparing for AI integration in research, Behzod Sirjani, Founder at Yet Another Studio, suggests diving in head-first with AI tools and gaining personal user experience: “Stop talking and go see it firsthand. I think that’s going to help all of us have a stronger perspective, and build better tools.”
If you haven’t already, bring AI into your daily toolkit and share them with your team. From data synthesis to AI and UX design tools, there’s some excellent tools already on the market. Get familiar with them, and keep an eye on new trends emerging in AI technology related to your field.
Enhance the human side of research
Eniola Abioye, Lead User Experience Researcher at Meta, says the addition of AI to workflows leaves a clear gap to be filled: AI cannot replace the “human element of research, because AI can’t do empathy, and it can’t do strategy.”
This means researchers have an opportunity to be recognized as a key strategic function in their organization. Researchers must take this chance to shift gears, and adopt a strategic approach to what is researched, and what action is taken—helping inform big-picture decisions about product focus and UX roadmaps.
Sri Putrevu, Senior Director of User Research at Circle, shares that "There's a real opportunity for researchers to embrace AI because it means we can scale. There's an opportunity for us to provide insights in-the-moment when decisions are being made, because we can now be in every meeting, we can be working with every engineer or product manager."
Ultimately, introducing AI to UX research isn't about replacing human researchers, but augmenting their skills. With AI’s potential to automate mundane tasks, researchers can embrace more tactical, strategic, and human approaches.
As researchers, we build empathy within the organization from observations made in the field.
Sri Putrevu, Senior Director, User Research at Circle
Sri shares that "Whenever I do research, one of the things that I'm always looking for is the emotion. How are decisions being made that are driven by emotions, subconsciously or consciously?" Understanding the emotions, behaviors, and experiences of users is a vital aspect of research—and one that AI cannot fully replicate.
Instead, AI research tools will free up humans to focus on this work that requires creativity, empathy, and strategic thinking—such as interviewing users, and extracting how those insights should inform the product—ultimately re-centering research and discovery onto understanding human motivation and needs.
Look for new opportunities
Like any emerging technology, many aspects of leading AI tools are still far from user-friendly. Whether it’s the unrefined UI or a complex, prompt-driven interface—for such advanced tools, there are still UX developments needed.
As Linus explains, AI tools remain an imperfect technology: “Sometimes they may work eight out of 10 times, but the other two times it does something quite unexpected.” While this isn’t surprising for such a complex and early-stage technology, misfiring AI can lead to a plethora of incorrect outcomes or confused users.
We’ve all wasted time arguing with a support chatbot that doesn’t understand the nuance of your query, or struggling to get voice recognition to actually recognize your voice. “During those times, the users can feel like ‘Maybe I didn't communicate properly? Is it my fault? Or is it that the technology's not working well?’” says Linus.
Part of the solution is educating users and guiding their interactions with AI. But a larger portion of the answer comes down to molding and refining the technology’s UX. With these challenges comes a new opportunity for UX professionals to fulfill a much-needed role, and help shape this emerging technology from the ground up.
Full speed ahead
Whether you’re first in line to try a new AI tool or are still on the fence, it’s clear that AI isn’t going anywhere.
As AI becomes more central to our daily experiences, the industry is only at the beginning of its journey with artificial intelligence. If we take a long-term view, AI in UX is still in its early days, so there’s many opportunities for research and UX professionals to experiment with innovative ideas, and bring about new developments.
Product teams are now presented with a landscape of research and product development where the winning organizations of tomorrow are the ones embracing the opportunities AI presents.
I think it’s a great time to be in UX; as the role will continue to evolve alongside this new technology, with user needs becoming more important than ever when building with AI or building AI tools.
While there is undoubtedly caution from some UX professionals, and some barriers to yet overcome, the benefits and opportunities far outweigh the uncertainty. With AI as a virtual research companion, product teams can see the potential to leap ahead of competitors in the new time-to-right.
UX researchers who embrace AI are set to evolve into an indispensable strategic pillar of their organization—one where insights are accelerated, and user experience becomes the central marker of a product’s success.