Can AI help us avoid bias as UX designers?
As a prospective UX designer that has attended interesting lectures about how to avoid biases, I have been taught about what bias really is as well as techniques to avoid it, but I still find it hard sometimes to let go of my own biases when conducting research.
It has become clear to me that as a designer it is important to understand what impact I have on my work, and how avoiding biases can help improve my designs.
But I am wondering, will AI be able to help us avoid bias in the future?
Let’s investigate this further, starting by quickly understanding what bias is and how AI works.
Bias
So, we humans are inherently biased, we have preferences and opinions. Bias means that a person prefers an idea or a thing over another, and may not give the other option an equal chance.
Now this is of course natural and may have many reasons behind it, such as experiences and feelings for example.
This might not always be a big deal, like deciding which type of soda tastes better, but when it comes to designing good user experiences it is important to lay our own preconceptions aside to make sure we have the user in focus and not ourselves.
Artificial Intelligence
Siri, Alexa and Google assistant are forms of artificial intelligence that most of us are familiar with, but AI extends far past voice activated assistance. AI is the simulation of human intelligence processes by machines. It works by giving your computer information and letting it make its own decisions and predictions about this subject. That is called machine learning.
And as Cassie Kozyrkov, Chief Decision Intelligence Engineer at Google Inc, wrote;
“To make matters worse, the whole point of AI is to let you explain your wishes to a computer using examples (data!) instead of instructions. Which examples? Hey, that depends on what you’re trying to teach your system to do. Datasets are like textbooks for your student to learn from. Guess what? Textbooks have authors.”
So that once again leads us back to biases. It seems like we are not able to get entirely free from bias, even when using AI, since it is us humans who teach the computer to think.
But there is actually a Swedish company that has come pretty close when it comes to using AI to minimize bias. Furhat Robotics have with their robot Tengai Unbiased been able to train the robot to conduct job interviews in a way that is very similar to how a human recruiter would do, asking competency-based questions such as “tell me about how you handled a conflict at your last job”. The robot will nod, smile and say “mhmm”, just like a human, and encourage the candidate to elaborate further on their answers.
After the interview, the robot will make a summary of the interview with an objective recommendation for a human to make the decision about the candidate.
With this in mind it does not seem unlikely that we in a few years also could have robots trained to conduct qualitative research for us, and hell, why not even compile the results too?
This could actually be a possible way to avoid confirmation bias, and I don’t think it’s too far-fetched!
My only concern would be that we miss out on the human touch, even though the Tengai robot looks pretty cute, I fear that we would lose engagement from the interviewees when they are making conversation with a robot, it is just not the same as having an interview over a croissant and a café latte with another human being that you somewhat can relate too.
Furthermore I believe that there are some follow up questions that could be ground breaking for your research that only you with your wicked human brain would come up with, and I doubt that the complexity, and actually even bias of that question is hard to simulate through an AI.
But then again, what do I know, I am not an engineer, I am just an aspiring UX designer who hopelessly believes her brain is better than any AI.
To sum up, everything is biased and it will continue to be that way, since we are all humans, but I’m positive that AI might be able to help us get rid of that extra unnecessary layer of bias, and will help us create better user experiences for everyone!
