Thread by @lizadixon

Reading a study about Tesla Autopilot + trust

There are some unreal quotes in here taken directly from @TeslaMotorsClub forums, including descriptions of repeated poor performance under familiar conditions which have led to passenger injury in other instances

Some examples👇

@TeslaMotorsClub Autopilot is an advanced driver assistance system. It's normal & expected that it struggles on-road under certain conditions...which is why it requires full-time, driver supervision and hands on the wheel

Promoting it as anything more is #autonowashing

@TeslaMotorsClub Another ex of Autopilot's functional limitations: T-Junctions, a common intersection style in which multiple accidents involving AP have occurred

miamiherald.com/news/local/com… *Driver: "Cruise Control was on." 95% certain CC=AP

+Here, an identical scenario is documented in a study:

@TeslaMotorsClub 🚨Warning: The link above contains graphic video of a T-Junction accident scene, in which one person was killed and another was seriously wounded.

Here's a link to the related study in the ss: researchgate.net/publication/32…

@TeslaMotorsClub These are not strictly Tesla AP problems. Many of these things could easily happen with other ADAS on-road

The difference? Tesla regularly refers to L2 Autopilot as if it is more capable than it really is (#autonowashing) + chose to forgo reasonable DMS

& so, Tesla gets extra🔬

@TeslaMotorsClub Also, consider: AP receives OTA updates. What affect might this be having on driver behavior?

Ex: You're aware that AP was updated last night. You're now on the highway. You want to "test" the new software, to see if an area in which it failed before is now "fixed"

WWYD?

@TeslaMotorsClub I really believe the above is not discussed enough.

There are also "soft" influencing factors that I think might lead AP users to push the system:

- Tesla culture, belief of bettering/saving mankind
- A personality type that may be drawn to the brand

& ofc #autonowashing