Tesla “Autopilot” “Safety Score” is Creating Better Kamikaze Drivers

Michael DeKort
4 min readSep 25, 2021

Here is a link to the Tesla process — https://www.tesla.com/support/safety-score

There are five areas of this test: Forward Collision Warnings per 1,000 Miles, Hard Braking, Aggressive Turning, Unsafe Following and Forced Autopilot Disengagement. Separate of “Autopilot” use the first four have value. However, they are being used here for mostly sinister and devious purposes, not to make these drivers or the public safer. Yes, they are meant to discover unsafe human drivers but primarily for the purposes of allowing them to be human Guinea pigs and to be more effective ones to test the next version of “Autopilot”. (If it were meant to make all drivers safer the “autopilot” software would be engaged in all the cars for this purpose. Not just those who pay $10k for “Autopilot”.) Once the “safe” drivers are discovered and their ability to commit Hari-kari is discovered or honed they now have the honor of being sacrificed as human Guinea pigs to train these systems to stop failing EVERY ONE OF THESE TESTS. To that the system is rigged to force that to occur through those very same failures.

Let’s look at the negligent and deadly irony and hypocrisy of a few of these tests

Forward Collision Warning

This test if fine assuming the object is moving inline with the car. If it is not and it is crossing or stationary the system will often run right into it because AP and AEB often fails here. Hence the two NHTSA investigations. One for all vehicles hit and the latest for emergency vehicles. This root cause has led to at least 5 fatalities.

Hard Braking

The irony here is brutal. “Autopilot” and Automatic Emergency Braking (AEB) are likely to fail when there is a stationary or crossing object in its path. Therefore, that system can’t hard brake in those instances because it doesn’t brake at all. That makes this test another false confidence misleading reuse. Why is this “test” here? Let’s look at the text. “Engage the brake pedal early when slowing down, coming to a stop or reacting to a change in your environment”. If the “safety driver” brakes too early the crash scenario, even the very beginning of it, is never experienced. Therefore, it cannot be tested and learned. This test is conditioning the human Guinea pigs to be a better Kamikaze driver.

Aggressive Turning

“Autopilot” currently struggles with lane boundaries. Including when there are clear lines. The system often drives into an oncoming lane. Elon doesn’t want the human driver to do this so they can remain alive long enough to sacrifice themselves so the system to fail this test so that it can learn from it. (And keep in mind, like all scenarios, this process takes massive repetition.)

Forced Autopilot Disengagement

This one has no ancillary positive byproduct for human driving. It is also the only test run only when “Autopilot” is engaged. It’s sole purpose is to mislead the world and provide false confidence. The issue is Tesla’s DM alarm has a delay of 20 seconds or more. How far does a car go in 20 seconds at 40mph? 1200ft or a quarter of a mile. In most cases crashes will occur well within that time frame. Making this warning and alarm a misleading, negligent joke. The “test” is making sure you don’t allow the reckless alarm time of 20 seconds to go to 21 seconds or longer. (In many cases the system has gone 30 seconds or more itself.) To fail these tests the human driver would have to be lucky to not have a crash in the alarm delay time frame. As such this process is not meant to avoid a crash it is meant to provide false confidence and legal air cover for those who can’t so simple math.

Clear Reuse

The final question is, will the world, especially Tesla Guinea pigs and regulators, see through this reuse? Will they see this is meant to mislead them, provide false confidence, make things worse, facilitate more crashes into emergency vehicles and kill people?

Update — Same Day

Update — Same Day

· There are now reports growing on Twitter that people are running red lights to avoid a Hard Braking issue and bad score. Shame running red lights is not part of the test?

· I made a mistake on the FCW alarm time and conflated it with the DM alarm time. I removed the FCW section and amplified the Forced Autopilot Disengagement section.

More on my POV here. Including how to do this right.

NHTSA should impose an immediate “Autopilot” moratorium and report initial investigation findings in 30 days

· https://imispgh.medium.com/nhtsa-should-impose-an-immediate-autopilot-moratorium-and-report-initial-investigation-findings-de5b6da4d704

NTSB Chair falsely states “We have done all we can do” with regard to Tesla’s “Autopilot” debacle

· https://imispgh.medium.com/ntsb-chair-falsely-states-we-have-done-all-we-can-do-with-regard-to-teslas-autopilot-debacle-be6c5e45f0f

The Autonomous Vehicle Industry can be Saved by doing the Opposite of what is being done now to create this technology

· https://medium.com/@imispgh/the-autonomous-vehicle-industry-can-be-saved-by-doing-the-opposite-of-what-is-being-done-now-b4e5c6ae9237

SAE Autonomous Vehicle Engineering Magazine — Simulation’s Next Generation (featuring Dactle)

· https://www.sae.org/news/2020/08/new-gen-av-simulation

How the failed Iranian hostage rescue in 1980 can save the Autonomous Vehicle industry

· https://imispgh.medium.com/how-the-failed-iranian-hostage-rescue-in-1980-can-save-the-autonomous-vehicle-industry-be76238dea36

USDOT introduces VOICES Proof of Concept for Autonomous Vehicle Industry-A Paradigm Shift?

· https://imispgh.medium.com/usdot-introduces-voices-proof-of-concept-for-autonomous-vehicle-industry-a-paradigm-shift-87a12aa1bc3a

Tesla “autopilot” development effort needs to be stopped and people held accountable

· https://medium.com/@imispgh/tesla-autopilot-development-effort-needs-to-be-stopped-and-people-arrested-f280229d2284

--

--

Michael DeKort

Non-Tribal Truth Seeker-IEEE Barus Ethics Award/9–11 Whistleblower-Aerospace/DoD Systems Engineer/Member SAE Autonomy and eVTOL development V&V & Simulation