Yes it should be regulated.
Yes it should be regularly inspected and maintained.

But where are the goalposts set in relation to "maximum level of safety" in regards to automobiles? Zero accidents ever, or is there an acceptable threshold?

Would reducing the number of passenger vehicle accidents by 5 million a year thus reducing the number of injuries by 1 million and dead by 22 thousand annually? Would this be a level of safety worth achieving, or should we put off deployment until the Maximum Level of Safety reaches zero?

Nevertheless, the performance standard needs to be at a higher level than it currently is with Tesla's Autopilot..

The Tesla has an insufficient set of sensors, and insufficient processing, to do what it CLAIMS to be able to do safely.

I've been watching a number of videos on autopilot. Even with v8.0, the system is making some dangerous, and potentially deadly errors. For example, the sonar blind spot sensors will see cars and busses, but not the gap between wheels on a semi. OK, IT thinks it is safe to merge right. And yes, it will feed you onto the exit ramp - AT FREEWAY SPEED.

Sure, you can talk all day about how people should know better. But unless and until there is some advanced testing to make sure unqualified and/or reckless people are not using the system, there is an obligation on the part of corporations who are intending to profit on the technology to use the safest system available.

For me to even remotely trust an auto pilot system, it has to have many years of great results and safety records of wide use.