Item — HN Reader

Here's on litmus test -- if it wouldn't be criminal/negligence for one kid to say this to another kid, should it be criminal/negligence for a chatbot to say it?

Like on the one hand, I obviously think that if this is a frequent behavior on commonly used chatbots, yes that's a problem. But if you have tens of millions of people talking to chatbots, I think it's almost inevitable that somebody is going to take something the bot says as a justification to do something wrong/illegal (in the same way that some crazy person is going to say a celebrity is in love with them, or that some crazy person will use a kitchen knife as a murder implement).

For now I think the world is a much better place with more competition in the AI market rather than fear of a lawsuit stifling smaller players.

3 comments

> if it wouldn't be criminal/negligence for one kid to say this to another kid, should it be criminal/negligence for a chatbot to say it?

Obviously it should be legal for billions of dollars in capital to behave -- at scale and for profit -- the same as children on a playground.

What a weird place this is.

If a doctor/therapist/adult said it it might indeed be criminal, and that is what the chatbot is replacing, not a fellow child companion.

> For now I think the world is a much better place with more competition in the AI market rather than fear of a lawsuit stifling smaller players.

This is essentially an arms/race, prisoner's dilemma situation. We need more players to evolve AI to a sufficient level, putting in more resources. The ideal would be to stop all chatbots and outlaw all completely, but typical economic thinking won't let that happen. The Amish don't have that problem because they have an oversight committee for each Amish community and they would outright reject such technology immediately.

Sadly, we do not have the wisdom and are thus letting this prisoner's dilemma play out to our collective detriment.