How many people will die in a FSD incident during the month of this trial?

A Tesla driver was arrested for vehicular homicide after he ran over a motorcyclist while driving using Autopilot without paying attention.
The accident happened in Snohomish county, Washington, on April 19.

@TychonNotos Very sad. 😭

Based on Elon's tweet, the trial was for a month, so I guess it's ending this week?

bought Ṁ20 Other YES

According to https://insideevs.com/news/629094/tesla-how-many-buy-fsd/, 19% of drivers have bought FSD, so this will increase the number of FSD on the road by 5x. According to https://www.cnbc.com/2023/08/25/tesla-autopilot-safety-probe-by-nhtsa-nearing-completion.html there were 26 crashes with fatalities from Aug 2019 - Jul 2023, about 0.5 per month. If we increase that by 5x, and take into account that all of the new drivers will be unfamiliar with the system, I think there's a good chance we'll be looking at 2+ deaths.

How will you be counting? Total fatalities of all incidents involving a FSD-capable model, or only ones where FSD was confirmed to be in use, or only where FSD was confirmed to at fault?

Or perhaps by calculating "excess deaths" relative to some baseline?

@retr0id Considering the bias implicit in the question, I bet the first one. However, excess death is the reasonable one.

@admissions it's likely that this trial will actually result in a net decrease in lives lost, but that's not measurable. This market will be decided by the number of people reported to have died in incidents where fsd was being used (even if caused by human error)

I have rephrased the question. Let me know if you want me to resolve NA and resubmit.

@GordanKnott Could you please elaborate on which outlet you will use for “reported deaths” (like Tesla, news sites, NTSB etc)