1. A few thoughts about the tragic incident in Arizona, where a woman was killed by an #Uber self-driving car…
2. Much of what I read in the first day after the incident clearly tried to shift blame to the victim, and to devalue that individual…
3. She was crossing outside a marked crosswalk…
She came out of the shadows…
She may have been homeless…
She had a criminal record…
She came out of the shadows…
She may have been homeless…
She had a criminal record…
4. Don’t question the tech – don't question its inevitability – question the human victim.
5. This is the section of road where the car, travelling north, hit the victim, who was crossing west to east. (image by @nytimes) 
6. The Tempe, Arizona police chief was quoted saying, “she came from the shadows right into the roadway”.
7. Yet photos clearly indicate that the victim was hit by the vehicles front-right corner.
The victim was crossing left to right.
The victim was crossing left to right.
8. And various media reports indicate the incident took place well into the roadway, not in the leftmost lane…
9. So a woman traveling west to east – left to right – progressed far enough into to road to be struck by the AV's right bumper…
10. …a 49-year old woman pushing a bicycle laden with shopping bags…
11. …yet the self-driving vehicle – based on current reports – did not stop, swerve, or slow down.
It either couldn't identify or respond.
It either couldn't identify or respond.
12. I’m not saying a human driver wouldn’t have hit her, but I wonder if that might be the case…or if the human would have at least braked.
13. If this is how they react to unexpected lateral movement into the lane, good luck to AVs travelling at high speed in moose/deer country!
14. Or, less facetiously, in any dense, complex urban environment with active roadside uses, and share ise of the right of way.
15. While AVs may never blink or fall asleep, they don’t understand a damn thing. They are rule-bound. Rules don't begin to capture reality.
16. Yes, as a human driver, my senses/attention are fallible – but I have intuition, and am able to fill in the big picture using cues.
17. My driving is the product of lived experience, of anticipation, of empathy, and – good lord – of an actual fear of hurting someone…
18. I am aware of the cat or squirrel by the side of the road, on its haunches, possibly poised to make a dash.
19. I am aware of the toddler, standing on the corner, tugging at her parent’s hand, a little too eager to cross…
20. I am aware of the body language of the person timing a dash across the street, trying to catch a bus…
21. I am aware of the ball that rolled into the street, and that there may be a child following close behind it…
22. Yes, humans rush, are emotional, are fallible - but how do you teach intuition? In complex space, with our behaviour, you need that.
23. An AV understands none of that – if you don’t fit its rules and model, you better hope you don’t wind up standing in front of it.
24. Think about a neighbourhood like Kensington or Yorkville – narrow streets, mix of users, pedestrians crossing freely…
25. Think about all of our buzzwords – animated, complete, vibrant, pedestrian-oriented – all antithetical to the smooth operation of AVs.
26. The types of complex, layered urban environments we most enjoy are those least well-suited to a world with autonomous vehicles.
27. My fear would be that instead of waiting for AVs to get it right, we are pushed to alter cityscapes to suit their limitations.
28. My fear is that ‘smart’ cars will require dumb cities. Sterilized urbanity that removes the nasty unpredictability of human behavior.
29. This tweet speaks to concern that corporate interests re: liability will ⬆️ pressure for ‘predictable’ streets.
30. And for what? Very possibly not to reduce traffic – a car on the road is a car on the road, however operated.
31. If autonomous vehicles are shuttling around empty after/between passengers, it’s entirely conceivable that they add to traffic.
32. If the point is to allow us to watch Netflix while moving – or to provide one more means for our data to be collected – at what cost?
33. Do they take our focus off of expanding transit, building bike lanes, improving walkability, and updating land use to shorten commutes?
34. The operating model of big tech is ‘move fast and break things’. Get past the point of no return before we know/consider implications.
35. Central to this is projecting inevitability. Don’t ask whether this should happen – just accept that it will.
36. I'm not really a Luddite, but I think change needs to advance - not hurt - the public interest. On our terms.
37. That last few days are a reminder to push back. Question inevitability. Challenge Uber. Challenge @SidewalkToronto. Challenge Facebook.
38. When an incident like this breaks the spell, we need to ask questions. Do we want this? What does it imply? What are the alternatives?
39. And if this is our future, how do we fit the tech to the world we want, rather than shape the world according to what the tech wants?
- End -
should add: I'm very sceptical that a 'safety driver' can effectively go from observing to taking control in crisis.
A few final thoughts: In 2016, there were 483 traffic fatalities in Ontario. 64% of those occurred on OPP-patrolled roads, largely highways.
I could see a case where you hand over control to AV tech on highways - eliminate the speeders, tailgaters, rubberneckers, and weavers...
...but if we are truly concerned about deaths in cities, we can lower speed limits and better design streets RIGHT NOW. #VisionZero
#VisionZero doesn't require tech - it requires compassion, foresight, and political will...
...and it doesn't demand reconfiguring urban environments for tech, or entrenching private motor vehicle dependence for another century.
Because as horrific the toll of traffic fatalities, the public health cost of sedentary auto-oriented lifestyles is likely worse.
So let's not tackle one public health problem by entrenching another.
- End -
- End -
