Human-centered design demands a limited use of AI-generated personas, if at all
A big debate among UX researchers is whether to use synthetic users in their studies. There’s one group of researchers who consider it no less than an abomination, unethical to use an artificial persona to test a product that serves humans. The second group defends it as a helpful tool that makes user research more efficient.
Looking at it through a pragmatic lens, I understand both perspectives. UX teams may consider using synthetic users for low-level artifacts, the parts of a study you can reasonably simulate.
Still, the word synthetic already tells you what you’re working with. These users aren’t real. As researchers and designers, our goal is to build products that solve real people’s problems. We need human feedback.
What if synthetic users can help us solve those problems without losing touch with the humans we design for? How far do we let them take us?
The problem of overoptimistic LLMs
A major challenge for any UX researcher trying to create a synthetic user is trusting a system that seems to always give them the answer they want to hear.
LLMs are way too positive and tend to flatter the person prompting them. Many companies are using this “prompt-to-please” research to their advantage, knowing the risks it entails. It’s simply too convenient to have a system that generates test users who love your product and only give positive feedback.
This, of course, creates a false sense of satisfaction that’s by all means unrealistic.
A 2025 study showed how this overly positive tendency of LLMs has affected UX research. Its title is a participant’s verdict: “She was useful, but a bit too optimistic.”
To conduct this study, the researchers gave eight UX designers an interactive AI persona called Alice to work with across three activities: user research, ideation, and prototype evaluation. Alice was fast and helpful, but she agreed with almost everything. The study describes “a consistent tendency to agree uncritically with design proposals.”
For example, a designer asked Alice whether a two-hour session about organic farming in her restaurant would help her gain customers. “Yeah, that could be really effective,” she answered. No doubts or constraints.
The researchers saw the same across proposals: responses stayed positive while the real-world complexities (cost, user adoption, resources needed) went unquestioned. Senior designers described the conversations as “one-way” and “self-affirming,” noting that Alice never treated a single idea as infeasible.
Participants did find her useful for early exploration. This is the use case I’d recommend UX design and research teams test first. A synthetic user can work well as a first touchpoint, as it can give you a broad understanding of the problem space.
Always keep in mind that over-optimism is persistent. LLM’s will almost always try to please the person prompting. We must explicitly ask for critique to get anything close to honest feedback.
Speed in preliminary stages of UX Research is a plus
Putting the tendency toward over-optimism aside, synthetic users do have one clear advantage: getting things done faster (at least in the early stages of the user research process).
It’s not easy to find a positive side to AI-generated personas, given that most opinions and articles are negative and focus on their limitations. However, I found a report showing that UX researchers do see some advantages, including speeding up their work.
User Interviews surveyed 150 researchers in May 2026 for The State of Synthetic Users study. The researchers who use synthetic users summed up their value in one word: convenience. Or, as stated in the report, “the ability to gather rapid feedback without the logistical pains of recruiting participants.”
Recruiting is the bottleneck in almost every study I’ve seen. In particular, finding the right people and scheduling interviews with them. Using synthetic users doesn’t require this step, which is one of the reasons why 41% of the researchers surveyed treat them as a preliminary tool, right before human research begins.
The methods where they’re rated most useful were survey and screener design (46%), usability testing for directional signals (34%) and early-stage research (32%). Nobody is using synthetic users to make the final calls.
There’s another, rather unusual application of personas that flips this idea around. I shared an example of it in one of my first newsletter issues this year, after reading an article by UX Writer Chinwe Uzegbu.
In that piece, Uzegbu describes how she used classic UX personas to improve AI, instead of using AI to generate personas. Her team was writing training queries for AI agents without knowing who the users were, so she built research-backed personas, including neurodivergent users, people with low technical literacy, blue-collar workers, and non-native English speakers.
The result Uzegbu obtained was richer queries that captured how people talk to AI. If that practice becomes standard, our personas can end up shaping the models that later generate synthetic users. Better models mean better synthetic users. Whatever level of quality synthetic users reach, they’ll owe it to properly conducted UX research by humans first.
How feasible is it to use AI-generated personas for UX Research today?
Conducting UX research successfully using only synthetic users isn’t possible. The speed at which AI is being implemented in our work may make this take obsolete in a few years, but as of today, human-centered research is 100% necessary and delivers the best results. Human complexity (our tone, hesitation, facial expressions, body language, and behavior) is irreplaceable.
People have emotions, and every person is different from the next, even inside the same profession. Their cultural experiences and mental models are different. Put five doctors of the same specialty in the same study, and you’ll get five sets of feedback, because they’re five different people. A synthetic user, by contrast, is homogeneous.
I spoke with Eric Mahlstedt, Vice President, User Research & Insights at SAP, whose team had already run the test: they recreated several of our previous UX research studies with synthetic users, then compared the responses with what they’d heard from real human participants.
Mahlstedt found the synthetic users’ responses to be overly generic. They were plausible, flat regurgitations of anecdotes anyone could find online.
“It felt less like conducting research and more like a sophisticated internet search, presented in a way that creates the illusion of empathy,” he said. “There is no substitute for talking to the humans who will be affected by a decision, especially when it’s a high-impact or high-risk decision, which we dedicate our research capacity to.”
We both agreed that synthetic users may be valuable as a hypothesis-forming tool. If a team is starting from a blank slate, synthetic users could help inform an initial hypothesis or identify areas worth investigating. They can provide a jumping-off point for further research, but never give absolute or definitive results because they can’t replicate human experience.
Synthetic users should only be used to scale human UX research
So, are synthetic users really the future of user research? No. That’s my flat answer. To get real people to do real work, you need real feedback. And the word “synthetic” says it all: these are not real people.
Yes, synthetic users can be used to scale a study, but they can never be the only data point we rely on to build products that are going to be used by humans. That’s the ironic part: we’re designing products for humans, so getting the data only from non-humans simply doesn’t make sense.
My advice is to use synthetic users only to extend the scale of human UX research you’ve already done. They can be another tool in the research process, but when it comes to key product research that affects people, synthetic research cannot be the one source you rely on.
Arin Bhowmick (@arinbhowmick) is Chief Design Officer at SAP, based in San Francisco, California. The above article is personal and does not necessarily represent SAP’s positions, strategies or opinions.
Are synthetic users (really) the future of UX research? was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.
