Psychological insight in AI design and development

Engineers often assume that AI models, especially those delivered via API, don’t need UX or human factors.  Yet, my experience of conducting human machine teaming research reveals the opposite: AI experiences really do need input from a psychologist!

For example, I recently studied the first iteration of a model trained to offer suggestions to novices and experts.

From the outside, it seemed like a typical usability test, however, because we were assessing novel AI features, I had to observe the model’s behaviour just as closely as the people participating as they worked together to complete a task.

Indeed, when conducting human machine teaming studies, it can feel like there are two participants in the test (human and machine), both deserving my careful attention as they collaborate.  

It helps to have another researcher there, one with more insight into the inner machinations of the model, to help explain those “aha” moments when we uncover issues and opportunities for improvement from the perspective of both the person and the model.  

To create successful experiences, we also have to understand the behaviours and peculiarities of the models as well as understanding expectations and behaviours of the people who intend to use them.

Contrary to common assumptions, AI models require UX design and consideration of human factors to ensure effective human-AI interaction. 

I should have called this article “you need a psychologist on your team!”

So, if you would like to leverage psychological insights in the design and development of your AI, then please get in touch!

Leave a Reply

Your email address will not be published. Required fields are marked *