AI has a UX debt problem
There’s a big problem out there in the world of new AI tools, and it’s been coming up more and more in the news and social media: these tools are freaking hard to use!
I wrote about testing AI agents recently, and my biggest takeaway was not that I could work faster, smarter, or more effectively by setting up a meeting notetaker or email responder. It’s that I lost time just trying to set them up… and failing at it. My conclusion was that without a better user experience, most of the tools will not be on the market for very long.
I am a technology early adopter and power user. If I can’t get it to work, how do you expect others to? That’s the point of UX, and right now companies are firing designers and researchers daily.
The issue does not only apply to AI agents. Other AI-powered tools on the market are missing the mark in UX. A recently published paper Ironies of generative AI: understanding and mitigating productivity loss in human-AI interactions found a number of areas where using generative AI has not resulted in increased productivity, and in some cases has made people less productive. For example, programmers using GitHub Copilot described shifting their work to “proofreading tasks,” but stated that they were taking the same amount of time to complete their work. When it comes to accessibility, AI is falling short. Many AI widgets and overlays are not playing well with screen readers, and AI-driven assistive technologies are not meeting blind users’ needs. As a spouse of a blind person I can also attest this this one personally, and as more and more companies add AI to their websites, there is a chance we will lose ground on making the internet more usable for all.
These companies are failing to follow some of the most basic principles and practices of UX and design thinking… and if they don’t start now, they will fail entirely. In pursuit of putting out their tools as fast as they can, they are skipping user experience. In some cases they are going as far as skipping finding an actual problem to solve, and instead just applying “something-something-AI” to a tool that already exists. In most cases, what exists already works just fine without AI.
The UX debt is already so high, and getting higher every day.
If users can’t, well, use the product, there will be no product in the future. Companies need to start investing in UX today, whether that means bringing in research and design to improve existing problems, or starting their process with research and design in mind. The longer they wait, the more costly that decision will be.
