5 ms·

As my old AI lecturer put it "If a person driving a car runs over a person that driver is at fault, if an AI driving a car runs over a person that AI will proba

As my old AI lecturer put it "If a person driving a car runs over a person that driver is at fault, if an AI driving a car runs over a person that AI will probably never be used again in any car.", so it's interesting to see the 'cute' approach they are taking to, as the article says "minimize the fear".

andrey-p 12y ago
I looked at the images and my initial gut reaction was "aww, I can't be afraid of _that_!" even if I still find the idea scary. I think it's a very successful design.
- ehurrell 12y ago
  I'd agree, except that quote stuck with me!
bitJericho 12y ago
One software change could rule out that same accident ever happening again. You can't retrain every driver everytime an accident occurs, but you can reprogram every automated car.
- ehurrell 12y ago
  Rationally yes, it's probably easy to argue that point, but his point, as a huge believer in AI applications, was that emotionally the 'killer robot cars' sentiment would break a fragile trust, as against the implicit common humanity of human drivers.
  - bitJericho 12y ago
    I realize this, but I disagree with the conclusion. I think just like today we have scarred individuals that refuse to drive, ride, or even take left turns, the same thing will happen with AI cars. Some people won't use them, but vast majority will realize we're better off, or at least feel that way. Some say we may be better off going back to the forests, and I don't know if I disagree with that some days!
    - ehurrell 12y ago
      I can understand your point, but I think the biggest differentiating factor when it comes to AI is that for maybe the last 20-30 years the biggest cultural touchpoint for the term 'AI' was the concept of 'it going nuts and taking over the world, enslaving/killing humanity', so I expect that to play into the public perception a little bit. It could be as simple as 'AI has a brand problem Google are working on solving', once that's been eroded in some way I'd essentially agree with you, I just think it's a looming perception issue.

andrey-p 12y ago

I looked at the images and my initial gut reaction was "aww, I can't be afraid of _that_!" even if I still find the idea scary. I think it's a very successful design.

ehurrell 12y ago
I'd agree, except that quote stuck with me!

bitJericho 12y ago

One software change could rule out that same accident ever happening again. You can't retrain every driver everytime an accident occurs, but you can reprogram every automated car.

ehurrell 12y ago
Rationally yes, it's probably easy to argue that point, but his point, as a huge believer in AI applications, was that emotionally the 'killer robot cars' sentiment would break a fragile trust, as against the implicit common humanity of human drivers.
- bitJericho 12y ago
  I realize this, but I disagree with the conclusion. I think just like today we have scarred individuals that refuse to drive, ride, or even take left turns, the same thing will happen with AI cars. Some people won't use them, but vast majority will realize we're better off, or at least feel that way. Some say we may be better off going back to the forests, and I don't know if I disagree with that some days!
  - ehurrell 12y ago
    I can understand your point, but I think the biggest differentiating factor when it comes to AI is that for maybe the last 20-30 years the biggest cultural touchpoint for the term 'AI' was the concept of 'it going nuts and taking over the world, enslaving/killing humanity', so I expect that to play into the public perception a little bit. It could be as simple as 'AI has a brand problem Google are working on solving', once that's been eroded in some way I'd essentially agree with you, I just think it's a looming perception issue.