Names are important. Calling it "autopilot" implies that it does the driving for you, and people will take that at face value. It should be named in a way that accurately describes its abilities.

If it is actually just assistive, and isn't autopilot, then calling it such is incredibly dangerous. It'd be like calling homeopathic junk a "cancer medicine", which is illegal mind you.

"Autopilot" in an aircraft doesn't imply it flies for you. And in an aircraft, there are multiple levels.

"Autopilot" is no worse than "cruise control" in this regard.

“Autopilot” is worse, not because of what “autopilot” actually means in aviation, but because of what “autopilot” suggests to the car-buying public, who are, generally, not aviators.

What does the car-buying public think those folks in blue suits are doing at the front of the plane?

Giving each other handskys!

No, I think many people just don't like the term and so decided everyone is confused by it. They're not.

It's not self-driving, it's autopilot. That's the contrast they're drawing, and it actually IS clear to the car-buying public. A car is the second biggest single purchase you're going to make in your life, so people aren't going to /not/ figure out what the term means before buying the car.

If the market is confused about the term and it's not clear in practice - it's still not a good term to use.

I definitely think it's worse than cruise control as you suggest. Cruise control systems control your cruising speed. Autopilot has way too many asterisks right now.

Pilots require years of specialized training. Airplane autopilots have been around for decades and you aren't going to have an issue with unskilled pilots attempting to rely on them for things they can't handle based on a name alone.

Also, airplane autopilots do fly the plane for you (including take-off and landing). They can certainly do so more safely than Tesla's "autopilot" does. A lot of this has to do with the aviation environment being much more controlled.

This whole comment thread seems like a distraction in the form of a non sequitur, though. If Tesla is falsely claiming capabilities that their system does not actually possess right in the name, and that causes deaths, then they could easily be liable, and at a minimum they need to stop.

There is no evidence that Tesla owners are getting misled by the name.

This very article describes a situation where the name of the feature doesn't seem to have played any role in the accident.

This is a case where bad word choice in marketing could literally kill people.

Agreed, with a couple caveats:

1. Any system that abstracts away human control bears some fault in failures of the human-machine system. Perhaps Tesla can get out of this legally through some fine print, or even ethically because the driver bears the final responsibility in the system. But they ought not design a system that is pleasant and convenient to use irresponsibly, which is what they have here.

2. Autopilot is supposed to handle issues like this. The Ars article says:

> ...a Florida highway crash when the Tesla he was driving struck a tractor-trailer as the semi was crossing an intersection of a divided highway that did not have a traffic signal.

So what if the intersection is unmarked. So what if the semi pulled out in front of him. So what if he ignored warnings. Autopilot should have seen the hazard and braked.

That's also the fundamental problem with all existing forms of autonomous vehicle AI. Google was the furthest along with Level 4 and 5 autonomy but even they have backed off that development track as they couldn't muster the resources (a massive fleet collecting data) necessary to make the next leap in progress.

Anything less than Level 4 autonomy will require constant driver attention. While there is technically Level 3 autonomy, the premise of a system which can operate with the operator intervening within some "limited span of time" in an emergency is fundamentally flawed. The situations in which an AI will have to kick the decision making back to the human driver are inherently time sensitive. Any sort of time-sensitive response from the driver would have to occur in such a short window that one could never safely remove their attention from the driving process for any meaningful amount of time. Even a time window as short as 5 seconds is too long to react to a situation at highway speeds (the car would have traveled 440 feet if it's at 60mph).

My point is that this is an existential problem in the field of autonomous vehicles to a point that even the most ambitious players are admitting defeat. Most, if not all, the commonly discussed benefits of an autonomous car (all the things you can do with a comfy seat and an internet connection in the time one spent driving, also no worrying about DUIs) are unachievable with Level 0-3 technologies. They might make driving safer if they're implemented as a fail-safe for bad drivers but if they're allowed to take full command for any period of time, they are fundamentally unsafe and undesirable in terms of what the public is demanding from the promise of a truly driverless car.

I am going to disagree there. Tesla names its system autopilot. Uses the fact that autopilot that includes collision avoidance reduces accidents by 40%, and then claims that there system is safer than a human (assisted human safer than a human, genius!). Elon musk has a statement every month about self driving autonomous driverless cars. They even have the gall to say that new cars have "full self driving hardware", is proof of outrageous statements even required? And all this hype/misleading advertising is fine because Tesla showed a few warnings? And the driver didn't have sufficient reason to believe the car was capable of complete autonomy. A simple Google search will tell you that this driver wasn't the only one.