And here's an interesting discussion, with further links, what can happen when collisions get you in trouble even though you're innocent: https://news.ycombinator.com/item?id=28236102#28238071
Also worth a read:
https://www.kammenlaw.com/blog/2019/11/how-often-are-people-falsely-accused-of-child-porn/
https://www.jamescrawfordlaw.com/blog/false-accusations-child-pornography-do-first
This largely depends on a country's law of course. There's been a somewhat recent case in Europe (Germany?) where a father was falsely accused and lost pretty much everything before he was found innocent in court.
That's why we need reliable and explainable AI systems, not just for safety critical applications like autonomous cars, drones, robots, etc. but for report systems as well. In general, any AI that makes a decision where the output can be catastrophic. We've seen some very interesting talks at the last GTC from various researchers in the industry about this topic.
Also worth a read:
https://www.kammenlaw.com/blog/2019/11/how-often-are-people-falsely-accused-of-child-porn/
https://www.jamescrawfordlaw.com/blog/false-accusations-child-pornography-do-first
This largely depends on a country's law of course. There's been a somewhat recent case in Europe (Germany?) where a father was falsely accused and lost pretty much everything before he was found innocent in court.
That's why we need reliable and explainable AI systems, not just for safety critical applications like autonomous cars, drones, robots, etc. but for report systems as well. In general, any AI that makes a decision where the output can be catastrophic. We've seen some very interesting talks at the last GTC from various researchers in the industry about this topic.