If they cannot detect a difference in behavior, is there a need to detect what isnt causing it?

I.e. there are so many different psychoactive substances, that effect, rather than presence, seems to be the only 'fair' way to discriminate in a world where most people take at least 1 medication a week or more.

Serious questions: What are the actual dangers of driving while high? If we can somehow normalize the amounts, which is more dangerous to driving: Smoking weed or drinking alcohol?

We know that cannabis use prior to driving increases the risk of an accident.

Alcohol is probably worse.

Allowing people to research cannabis would let us know what the risk is.

How do we know that?

(I expect it is higher, but that seems hard to produce meaningful data. Self reported? Subject to a blood test? A piss test that detects 2 months use?)