The Impact of AI on UX research: Opportunities and Challenges
14 June 2024 - By N. Hoeberichts
The rise of AI models and tools is obviously a hot topic currently in many industries. Also in the field of UX research, it has opened both opportunities and challenges for UX researchers in their jobs.
We’re still coming to terms with how AI can exactly help out UX researchers in their jobs and how it might threaten it also.
In this article, let’s dive a bit deeper into the implications that new AI technology and tools have on the field of UX research. Specifically, we’ll look at opportunities and challenges for the UX research field.
Opportunities
Support in data analysis - both quantitative and qualitative
AI can be of great support when it comes to analysing UX Research results - both qualitative and quantitative. Its ability to sift through data in a matter of seconds and derive meaningful insights out of this data is quite amazing.
Many new services and tools are popping up that can help analyse large volumes of text to find underlying themes and patterns. Depending on the tool, you can easily select which topics and themes you want it to categorise the feedback into and there are often great options for visualisations of the data too.
Because thematic and qualitative research analysis can be quite time-consuming and a bit of a pain as you need to manually sort insights and find patterns in largely unstructured qualitative data, getting support from AI on this is extremely beneficial.
However, we can see that there are still gaps in the output that the AI gives us, as it may not understand all the context that is behind what a human says in a research study. But for simple text analysis, such as customer feedback, it can save a lot of time that is spent going through the feedback manually and coding it.
AI as a research assistant
ChatGPT has already proven itself to be a valuable assistant when it comes to important UX research activities such as writing and reviewing interview scripts, writing research proposals and giving advice on what kind of research to run depending on the research question or problem.
Quickly running your research proposal or research script via ChatGPT may uncover certain topics or areas that you may have missed, after all we’re all prone to human error. As such, using ChatGPT as a kind of virtual research assistant and even advisor can help tremendously, especially if you’re the only UX researcher in your organisation and you might not have anyone giving feedback to your research outputs.
Creating better user profiles and segments
Using AI to analyse all your customer and user information to create improved user profiles and segments is another opportunity that can definitely be exploited using AI.
AI tools can analyse vast amounts of data from different sources, such as user behaviour, browsing histories, and customer feedback, to create detailed and accurate user profiles. They can identify patterns and correlations that might be challenging for humans to spot, making it easier to do a segmentation of users into detailed persona profiles based on their behaviour, preferences, and demographics.
Predictive Analytics
Using AI, UX researchers are able to make better predictions on a user’s needs and behaviours and how it might change in the future, using past data to inform the AI model. It could also potentially identify usability issues before they even arise, helping out UX practitioners with creating a more user-friendly interface.
I'm fairly certain that at some point we will be able to upload a interface design and the AI will be able to give specific recommendations to improve usability for the relevant persona. I don't think we're there yet, but it is coming.
Challenges
Job Displacement
Perhaps the biggest challenge when it comes to the implications of AI on the field of UX research is the fear that it may put a lot of UX practitioners out of work. Simple UX research, such as validation usability tests, might be done with AI tools that can run unmoderated studies.
However, it will be hard for AI to replace more complex research studies such as contextual inquires and ethnographic studies. These kind of research methods require a lot of empathy and knowledge about human behaviours, and that is something an AI is not able to fully grasp, at least for the time being. Important human contexts, such as emotions and human biases, might be missed by the AI and will need a human researcher to correctly interpret the qualitative data.
Data security
There is also some concern when it comes to the ethics and security of using AI for UX research, depending on how you use it and what for of course.
For example, uploading large quantities of user feedback, along with personal identifiable information, to an AI platform on the cloud may not be the most secure way of doing things, especially when it concerns sensitive information (e.g. medical research results). People may not want whatever they said in an interview to be uploaded to an AI model, so proper consent might need to be requested at the start of the interview.
Data reliability
While AI can no doubt be very helpful in analysing research results, it has been shown to sometimes amplify the bias that is often found in analyses done by humans. Since the AI model has been trained by those humans in question, it will inevitably pick up on those biases too (e.g. confirmation bias). So, how can we make sure that it won’t include such biases when the AI is doing the analyses?
If we want to ensure data reliability when using AI tools for UX research, we will need to train AI tools better in recognising the human biases that may exist in the data we upload for the AI to analyse.
The Future: Collaborating with AI
We’re still in the midst of figuring things out when it comes to how to best collaborate and work with AI models and tools to get the most benefit out of it. And also in the field of UX research this is the case.
When it comes to using AI in qualitative and quantitative research analyses, we need to ensure that the process is transparent, that any biases are addressed and that it’s clear how the model made those extrapolations. This may mean training the AI to explain how it’s doing the analyses and letting it show us explicitly the steps it’s done when performing the analysis task.
Also in terms of data security, steps need to be taken to make sure no personal data is compromised and potentially misused by AI tools and the companies behind them.
I do believe that the advancements in AI technology can be of tremendous advantage for UX researchers and other UX practitioners - we just have to figure out how it can best support us in our tasks and how to solve the challenges we face when it comes to using AI in our UX research work.
I'm confident that we will be able to leverage AI models and tools for our benefit , helping us do time consuming tasks like research analysis faster and better than before.