Feels like DS hiring logic is starting to change because of AI

Been noticing new DS hiring products like Litmetrics.ai lately, which seems much more focused on real datasets and messy business cases than the classic coding-test format.

A lot of DS work today are more like to be end-to-end analytical judgment with AI in the loop. That feels like a different hiring target than the classic CodeSignal / HackerRank screening - pretty sure most DS have used them in interviews.

Curious what other people think. Is DS hiring actually changing on the assessment layer - to whether candidates can work through an real business problem, or putting AI language on top of the classic coding test & screening process is still the best way?