Are you willing to take that 5% chance in your next interaction with police? Why not 94%? Or 96%? Who defines accuracy? Who measures it?

Accuracy: dog alerted -> no drugs found -> dog was incorrect, hence inaccurate

I consider 95% to be highly accurate and reasonably acceptable, which is what I'd want in a drug dog. But I have no knowledge of what it should be, I'm not a dog expert. That's who I'd ask when making the determination of accuracy %. If it's anything less than 95%, though, I'd remove dogs completely.

If you are the subject of a traffic stop do you want to take a one-in-twenty chance that the cops find drugs on you, even if you do not have them?

95% is not highly accurate, it is arbitrary.

It is what I consider highly accurate and acceptable, and the chance I am willing to take. You are free to disagree. Just because the dog alerts doesn't immediately mean you are guilty.

The point of my comment is that there should be a standard at which the dogs have to meet.