From the even-a-stopped-clock-is-right-twice-a-day department:
FTC Chair Andrew Ferguson told Reuters that he thinks that companies should be liable for chatbot harms. At that time, Ferguson emphasized that he was tired of people anthropomorphizing AI tools. He insisted that, “if someone tells a tool to do something, and the tool does it, I don’t think we would say, ‘Oh, what do we do about the tool?’”
If you build a tool and in the process of building it it causes harm, that should be on you. If some model under development hacks into another company, the developer should be held criminally and civilly liable.
If you build a tool and someone else uses it to cause harm, that person is 100% liable. It is not your responsibility to prevent your tool from being able to cause every possible harm.
If you build a tool and someone else uses it as intended, but it malfunctions and causes harm, the developer should be liable for that just like any other product liability situation.
Think of it like a riding mower. If I bypass the safety systems and point it toward my neighbor’s house and it runs over their dog, that’s on me not the manufacturer. But if I am riding along and suddenly it knocks me off the top and keeps going on to the neighbor’s yard and runs over their dog, that is on the manufacturer for building an unsafe mower.