Imagine 100 self driving vehicles in Toronto or any other city on earth doing what the van did today there?

At present this would be feasible with present Self driving vehicle technology. And anyone programming it to do this could jump out or do this programming of the vehicle remotely

However, remember what I said about the more simple the software and hardware design the less likely there would be problems?

Well, in regard to hacking the opposite is true.

The more constraints the hardware and software on a self driving vehicle the less likely it could be hacked to do what the van did today in Toronto.

So, on the one hand just in regard to general safety you would want something as simple as a VW Bug in the 1960s and 1970s in regard to both software and hardware for the least amount of general problems.

But, in regard to hacking creating what we saw today of 9 killed and many many injured (in regard to a hypothetical self driving vehicle), theoretically someone could with enough self driving cars, vans and trucks and semis kill could kill  thousands  of people around the world on sidewalks on any given day at present with enough self driving vehicles.

So, the more constraints the more likely people will have general fatalities but there would be less likelihood of terrorists or attacks by the insane all over the world.

Also, remember if these 100s or 1000s of self driving vehicles were hacked remotely, likely there would be no one to be found (at least in this country) to blame for thousands of potential deaths either here or around the world.

So, this is an ongoing potential problem regarding ALL self driving vehicles.