Exactly this. Pot is probably the most common and it stays in your system for 30 days after use (probably the longest that is tested). Essentially the screen will fail a guy who smoked pot two weeks ago, but pass a guy who shot up heroin all last week. Like the polygraph test, they are deeply flawed, in this case, for testing for intoxication.
Also, imagine if drug tests failed an applicant for having a beer in the last 30 days. Like it or not, pot is replacing beer for a large portion of Americans, time to adapt.
https://equilibriabook.com/toc/
Here, I think it's something like, "insurers move very slowly to new systems, and until this new testing has a lot of data behind it, they won't adjust premium. to account for it being better. Furthermore, juries will continue to give punishingly big damage awards for using it, on the mistaken believe that a toker shouldn't be on the job even if he wasn't high and passed the reaction test."
Better is to use impairment testing. An impairment test is a test of cognitive function, alertness, reaction time, hand-eye coordination and similar objective things that actually correlate well with how well you can do your job now, instead of sometime in the past.
On top of that, impairment testing can catch when you are impaired for reasons OTHER than intoxication. If your aircraft mechanic is not fit to do maintenance on your jet that morning because last night he found his wife sleeping with his best friend, they argued all night, and she announced that morning she is getting a divorce, leaving him a tired, emotional wreck who will make mistakes left and right at work...an impairment test can catch it. A drug test cannot.
In addition, generally there is lag between drug testing and results. Impairment testing can be quick (a few minutes) and cheap so that you can do it to every employee, every day at the start of their shift.