The tooling isn't there
A lot of testing tools were quick to jump on the AI bandwagon, it was disingenuous and was limited to self healing selectors in a DOM. Big deal, you still have to maintain those tests... fudged AI buzzword tooling or not.
Applitools and Leapwork might have some nifty tooling, again mostly for browser DOM and visual changes in a browser... but these are very expensive platforms (Applitools is anyway, not so sure on Leapwork), so forget about it unless you work for a big company with money to spare. Opening a browser and interacting with it is an expensive operation and can be a brittle way to do things but browsers do not equal LLMs.
I could use AI currently to generate some test cases perhaps, or send inputs via an API to verify the output... it's basically black box testing and it's basically slow.
Biting the hand that feeds IT
