Hacker News Reader

Comment by HeatrayEnjoyer | Hacker News Reader

The onus is on organizations to not sell or offer knowingly harmful services to the public. Any judge or jury will interpret a company directly advising the customer to jump to their death as knowingly harmful. Every engineer at an AI company is one dead person in a wealthy family away from legal liability, maybe criminally.

View on HN

I think interpreting a chatbot’s output as a company’s “direct advice” is an unsubstantiated jump. In the same way that if an unhandled error code in a program resulted in “DIE DIE DIE”, “Abort/kill child”, etc (trying to make these programming-related here) being alerted to the user, no reasonable person would hold the company responsible for the user taking this as a command. Imagine if we lived in a world where arbitrary text being sent by an entity could result in prosecution? I just don’t believe the chat UI to be a big enough bridge between the two examples to justify it.

A wealthy family? If justice is a contest of money, there are few families who can outbid the various parties interested in selling chatbots.

Even a poor family will get the backing of lobbying groups who's best interest is to ban their competitor.

Not to mention, the moral coating, which matters at least a tiny bit in domestic politics. (Note, I think the moral coating is actually incorrect, but the majority of people are not thinking critically, they see N=1 and make emotional decisions)

Reply on news.ycombinator.com