Member-only story
Dumb AI is pushing out good UX
A cautionary tale about the Washington Post.
Over the past four months, the Washington Post, my local newspaper, has doubled down on inserting poorly conceived AI throughout its website, to the detriment of good UX.
First came a revamped search feature that violates well-established best practices. Then came an AI-powered commenting system that has prompted hundreds of subscribers to complain vociferously on every story and then finally throw in the towel after being ignored.
If this is the future of AI for newspapers, we should all be worried.
The importance of well-designed search
Research shows that a good search function is crucial in engaging and retaining users, whether customers or subscribers. For example, a study by the Nielsen Norman Group found that users who successfully apply advanced search filters are more likely to find what they’re looking for and feel satisfied with their experience.
A well-designed search interface can significantly enhance user engagement and retention. — UX Planet
When I was a Senior UX Designer at PBS, I initiated an overhaul of the search feature on PBS LearningMedia to facilitate teachers being able to easily see how to filter results by facets of relevance to them. The revamp took search best practices* into account, including implementing auto-suggest and using filters and facets that our team’s research determined were of most relevance to users. The update focused especially on improving usability on mobile devices.
The Post’s AI-powered search
For years, the Post’s search results page mostly adhered to those best practice recommendations (see links in footnotes). Users could sort by date range, and filter by section of the newspaper and even by, if memory serves me, byline.
The Post’s new search function, unveiled late last year, casts aside best practices in favor of a minimally usable AI that tries to “answer a question” instead of finding articles relevant…
