Comment
Comment by delichon
An algorithm accidentally enticed a child to hang herself. I've got code running on dozens of websites that recommends articles to read based on user demographics. There's nothing in that code that would or could prevent an article about self-asphyxiation being recommended to a child. It just depends on the clients that use the software not posting that kind of content, people with similar demographics to the child not reading it, and a child who gets the recommendation not reading it and acting it out. If those assumptions fail should I or my employer be liable?