“I’m Sorry Dave, I can’t do that…”
So said Hal in ‘2001’, courtesy of Arthur C Clarke.
We’ve just returned from time away learning new things, seeing old friends, making new ones and reminding ourselves why de-camping to mainland Europe before 11pm 29th March 2019 might be a good idea.
On the way there someone was using ‘Automated Cruise Control’ on the motorways then the autobahn. Always entertaining as the car slows and speeds up for you.
There are a range of things that the car cannot ‘see’ which are listed in the instructions, with wording to the effect that even if you think it should have seen something you’re still the most sentient being present and all decisions are ultimately down to you. It’s next to the bit about not expecting the car to break the laws of physics.
So, as we were driving along, a car cut in from the side. So far so good. Our car ought to see that. And to an extent it did. It wrote, “Brake!” on the ‘message to driver’ bit of the dashboard then promptly switched off the automated cruise control. If it had been able to, it would have put its little wing mirrors over its headlights and whimpered.
Turns out that in order for the car to be able to brake by itself, it has to see both the obstacle and that there’s sufficient room for braking at a slow enough rate. Braking faster than this automatically is apparently too fast for humans involved and so the control is handed back to them. We braked in time with no difficulty, it was just a bit unexpected after the car had ‘seen’ everything else on the journey so far without issue. The picture in the manual shows that the sensor is in the middle at the front, meaning that anything coming directly from the sides wouldn’t be seen.
It reminded me of the reactive glasses tint – I have it on good authority that when first developed they were virtually instantaneous, but this didn’t allow time to mentally adjust and it scared the wearer. Now even though they can change faster they don’t.
This did prompt a discussion about Asimov’s laws of robotics, noting that it possibly went against all three, as a crunch would harm both the car and the humans.
Perhaps the car has developed the priorities of ‘I, robot’ VIKI, preserving the passengers by reminding them periodically to pay attention or persuading them to stay home.
Either that or the car has worked out a whole new set of priorities using a whole extra level of sentience we’re not giving it credit for, along the lines of Dave, especially if hypothetically a new car would not automatically be an insurance write-off (I’ve seen some scary repairs done on cars that should have been), but humans dent easily.