Philip Koopman on Autonocast–Ethical, Moral and Professional Chameleon
The Koopman Autonocast Episode — https://www.autonocast.com/blog/2019/5/3/142-dr-phillip-koopman-of-edge-case-research-on-safety-engineering-for-autonomous-vehicles
First, I would like to take a moment to readdress the positions I take and the direct nature of my advocacy of them. I am aware this approach can be off-putting. And given we are about to release our L5 scenario/simulation system product demo, I am also aware sales 101 does not suggest using that approach with potential customers. This is the reason I originally tried to work with simulation companies in the industry to help them fix significant gaming architecture induced capability gaps they had and to add scenarios and proper systems engineering practices to it. After these companies said they would not make the changes we showed them were necessary until their customers discovered them and I saw no one was offering any one piece with the fidelity and scale needed, let alone even attempting to create the integrated whole, I decided to create my own company.
Over time I can see that my efforts, and real-world events, are shifting the industry’s position for the better. Simulation has become the talk of the day. Problem is, especially given Elon Musk’s new L4 rampage, the paradigm shift probably won’t complete before there are more unnecessary deaths, especially the first woman or child. My hope is that potential customers come to understand my intentions and my methods are the right ones. That they come to realize that echo chambers, especially massive ones like this, don’t shift because someone says please. Like the aerospace industry begat the FAA, it usually takes multiple tragedies. I guarantee if any of the people in power in any of the companies making autonomous vehicles had loved ones who died because of public safety driving, especially to learn accident scenarios, they would feel very differently about my approach.
This now brings me to Philip Koopman. There are several “leaders” in this industry I am so disappointed in and frustrated with it has clearly, and quite unfortunately, become contempt. These are people who squander their pedigree, influence and power by choosing personal and professional gratification over doing the right thing. They are wolves in sheep’s clothing, Pied Pipers who lead people astray and many to harm needlessly. Either physically, emotionally or financially. Philip Koopman, Elon Musk and Mark Rosekind are examples of this. As with Musk and Rosekind, Koopman misleads people into thinking the use of public shadow and safety driving is the best or only way to create driverless vehicles. And the injuries and deaths that approach has and will cause are necessary to save more lives later. He also perpetuates the belief that simulation cannot replace most of that practice. He is of course, wrong on all counts. It is impossible to drive the one trillion miles or spend over $300B to stumble and restumble on all the scenarios necessary to complete the effort. Many of which are accident scenarios no one will want you to run once let alone thousands of times. Also, handover cannot be made safe for most complex scenarios, by any monitoring and notification system, because they cannot provide the time to regain proper situational awareness and do the right thing the right way. Said differently the industry and these three “leaders” are doing the exact opposite of their stated mission. They will never get close to L4/5. That means the lives that technology would save will not be saved. To make matters far worse the industry is taking lives needlessly as test subjects. With regard to simulation being able to replace most of public shadow and safety driving. Aerospace/DoD simulation technology, especially the type used for DoD simulated urban war games, is beyond capable of replacing 99.9% of public shadow and safety driving.
This set of quotes demonstrates exactly what I am talking about. Koopman said there should be a “road testing safety standard”. That statement means he supports public shadow and safety driving to develop and create autonomous vehicles. However, prior to this he stated you can’t just “drive around until you are safe” because it will take “billions of miles.” He also mentions using aerospace/DoD practices when he makes his case that AV makers need to make an argument for why they are safe. A year ago, he did not mention it is impossible to get to L4 due to the billions of equivalent miles nor aerospace/DoD practices. He does so now because he feels the wind shift and wants to keep his feet in both camps so he can drift to where the “conventional wisdom” and money is as opposed to having the courage to voice the right professional, ethical and moral position. The problem is he leaves the industry in a no-win situation. You must use public shadow and safety driving to do most of the development and testing but can never finish. And simulation, as he understands it, doesn’t fill the gap. At the end of the podcast he didn’t even have the courage or integrity to take a stand against Tesla’s promise to get to L4 this year and L5 next. Leaving the door open to morph into what Tesla wants to stay on the inside and make money anywhere he can.
What he should be saying is that no real-world safety driver testing should exist unless:
· The use cases or legitimate geofence being used has been fully tested in simulation and on test tracks first
· The AV maker proves no simulation technology, including using technology from aerospace/DoD, is sufficient before they ask to utilize test tracks or the real-world
· When the real-world is needed the events should be run in a controlled manner. Not unlike a movie set. This includes targeted and protected areas where living beings are not involved unnecessarily.
Full disclosure — Autonocast interviewed me a while back, refused to air the episode and will not explain why that decision was made. I believe it involves their belief that public shadow/safety driving is tenable, the lives it will take are for the greater good and simulation cannot replace it. Or that they have no interest in poking the bear that hard. (This is despite their never having seen the aerospace/DoD simulation technology I refer to. Nor accepting my offer to provide them proof. I believe this is in part due to Alex Roy being an investor in an AV simulation company that utilizes the gaming-based architectures I mentioned.) I also reached out to Philip Koopman’s company Edge Case a long time ago to see if they wanted to partner. This was before I knew Koopman’s positions on public shadow and safety driving, his windblown ethical compass or his being a poser. Additionally, I, an IEEE Barus Ethics award recipient, another IEEE Barus Ethics award recipient and Dr. David Cummings requested IEEE review whether Dr. Koopman’s efforts merited the IEEE Barus Ethics award he was recently presented. This because of his position supporting public shadow driving for autonomous vehicle development and testing and his efforts in the Toyota unintended acceleration case. Notably his expert witness testimony on unintended acceleration being disallowed in a case against Ford by a federal judge for not being adequate. A decision that was recently unanimously upheld by 3 federal judges on the Court of Appeals. (Ref link below) Finally, as I stated above, I have created a company called Dactle. We are producing a system that will utilize aerospace/DoD simulation technology and systems engineering practices to provide all the scenarios and simulation needed to get to a legitimate and verifiable L4/5.
Please find more on my POV below. If someone wants to see proof aerospace/DoD simulation technology can replace 99.9% of public shadow/safety driving and resolves significant issues with simulation technology based on gaming architectures, please let me know.
Matricksz.TV Podcast — Featured Guest
The Autonomous Vehicle Podcast — Featured Guest —
SAE Autonomous Vehicle Engineering Magazine-End Public Shadow Driving
Common Misconceptions about Aerospace/DoD/FAA Simulation for Autonomous Vehicles
Using the Real World is better than Proper Simulation for Autonomous Vehicle Development — NONSENSE
How Driverless Vehicle Makers Should Prove their Technology Works — —
The Hype of Geofencing for Autonomous Vehicles
Expert testimony dismissal link — pages 14–20 — https://law.justia.com/cases/federal/district-courts/west-virginia/wvsdce/3:2013cv06529/106000/1175/