Photo: Euro NCAP
Philip Koopman already said that Tesla is not using safety factors in its Safety Score Beta. What the company is adopting are risk factors used by insurance companies – which raised suspicion that the Tesla software had other goals than granting access to FSD (Full Self-Driving) Beta. Consumer Reports went even further: it tested the software and expressed concerns that the idea may increase dangerous behavior behind the wheel instead.
According to the consumer protection organization, Tesla drivers are discussing on Twitter how to increase their ratings with practices such as coasting at stop signs or accelerating through yellow lights. Stopping would penalize them for a 0.3 G hard braking. Consumer Reports measured what that force represents.
With the help of an accelerometer, the organization discovered that letting the car slow from 25 mph (40 kph) until it stops would already achieve a 0.3 G and hurt the Safety Score. If the car is on Autopilot, nothing that happens with the EV apart from forced disengagements of the system counts. In other words, Autopilot can control the vehicle dangerously according to the Safety Score Beta standards, but that will not impact the rates. Go figure…
Consumer Reports kept on testing stops, and only one at 0.5 G would look like a moderately hard braking event. Despite that, it would still not lock the seat belts, which shows Tesla’s standard for that may really lead to people avoid braking when they should.
Like autoevolution, Consumer Reports did not find which score is necessary for people to have access to FSD Beta, a feature for which some of them paid up to $10,000 in advance. However, a tweet from Elon Musk clarified that there might not be a minimum. The plan would just be to add around 1,000 new users per day.
As you may have guessed, the first ones to get access will be the ones with the higher scores. According to Musk, “first few days probably 100/100, then 99, 98, etc.” At the same time, the Tesla CEO said that the Safety Score Beta is what its name implies: beta software. That means “It will evolve over time to more accurately predict crash probability.” In the end, what you have is beta software granting access to beta software. What could go wrong, right?
For that to happen, Consumer Reports suggests that Tesla changes its metrics. According to Kelly Funkhouser, the organization’s head of automated and connected vehicle testing, the company should make use of the information its Tesla Vision system is able to retrieve. If it can identify bike users, pedestrians, cones, and emergency vehicles, braking for them should grant the driver points, not take them away.
Another change the organization believes the Safety Score Beta deserves to become a stable release software is determining if the driver is speeding. Tesla Vision and GPS data should be enough to assess if a driver respects the speed limits or not.
Finally, Consumer Reports urged Tesla to improve its driver monitoring system. After all, FSD Beta access should only be given to people that prove to be paying attention to the road at all times, regardless of engaging Autopilot and FSD or not. Three people have already died for overreliance on Tesla’s alleged Level 2 ADAS (advanced driver assistance system): Joshua Brown, Wei “Walter” Huang, and Steven Michael Hendrickson.
One way to measure safe driving would be for the Tesla app to evaluate if the driver uses the cell phone while driving, for example. As Jake Fisher put it, “if Tesla is trying to encourage safe driving, it should focus on whether the driver is paying attention.”
The senior director of Consumer Report’s auto test center makes a fundamental point. Ensuring the driver is paying attention is crucial not only for better regular driving. It is the most urgent task for people using FSD Beta, which Tesla itself said that “may do the wrong thing at the worst time.” If having 1,000 drivers more on the streets every day with that software – approved by another beta software – does not concern you, nothing will.
With the help of an accelerometer, the organization discovered that letting the car slow from 25 mph (40 kph) until it stops would already achieve a 0.3 G and hurt the Safety Score. If the car is on Autopilot, nothing that happens with the EV apart from forced disengagements of the system counts. In other words, Autopilot can control the vehicle dangerously according to the Safety Score Beta standards, but that will not impact the rates. Go figure…
Consumer Reports kept on testing stops, and only one at 0.5 G would look like a moderately hard braking event. Despite that, it would still not lock the seat belts, which shows Tesla’s standard for that may really lead to people avoid braking when they should.
Like autoevolution, Consumer Reports did not find which score is necessary for people to have access to FSD Beta, a feature for which some of them paid up to $10,000 in advance. However, a tweet from Elon Musk clarified that there might not be a minimum. The plan would just be to add around 1,000 new users per day.
As you may have guessed, the first ones to get access will be the ones with the higher scores. According to Musk, “first few days probably 100/100, then 99, 98, etc.” At the same time, the Tesla CEO said that the Safety Score Beta is what its name implies: beta software. That means “It will evolve over time to more accurately predict crash probability.” In the end, what you have is beta software granting access to beta software. What could go wrong, right?
For that to happen, Consumer Reports suggests that Tesla changes its metrics. According to Kelly Funkhouser, the organization’s head of automated and connected vehicle testing, the company should make use of the information its Tesla Vision system is able to retrieve. If it can identify bike users, pedestrians, cones, and emergency vehicles, braking for them should grant the driver points, not take them away.
Another change the organization believes the Safety Score Beta deserves to become a stable release software is determining if the driver is speeding. Tesla Vision and GPS data should be enough to assess if a driver respects the speed limits or not.
Finally, Consumer Reports urged Tesla to improve its driver monitoring system. After all, FSD Beta access should only be given to people that prove to be paying attention to the road at all times, regardless of engaging Autopilot and FSD or not. Three people have already died for overreliance on Tesla’s alleged Level 2 ADAS (advanced driver assistance system): Joshua Brown, Wei “Walter” Huang, and Steven Michael Hendrickson.
One way to measure safe driving would be for the Tesla app to evaluate if the driver uses the cell phone while driving, for example. As Jake Fisher put it, “if Tesla is trying to encourage safe driving, it should focus on whether the driver is paying attention.”
The senior director of Consumer Report’s auto test center makes a fundamental point. Ensuring the driver is paying attention is crucial not only for better regular driving. It is the most urgent task for people using FSD Beta, which Tesla itself said that “may do the wrong thing at the worst time.” If having 1,000 drivers more on the streets every day with that software – approved by another beta software – does not concern you, nothing will.
Wow, lot of interest in FSD beta! Plan is to roll out version 10.2 midnight Friday, then on-ramp ~1000 owners/day, prioritized by safety rating.
— Elon Musk (@elonmusk) September 28, 2021
First few days probably 100/100, then 99, 98, etc.
— Elon Musk (@elonmusk) September 28, 2021
Very much a beta calculation. It will evolve over time to more accurately predict crash probability.
— Elon Musk (@elonmusk) September 25, 2021
Google News