This is what LLM's do, they respond to user prompts. It's the user who is at fault, not the LLM. If I smash a glass with a hammer, does theregister publish an article blaming the manufacturer of said hammer?

Doesn't that create a dichotomy? On one hand AI is super intelligent, and on the other it must be treated like a small child.

Also, if you convince someone to commit a crime, do you bear no responsibility?

No.

AI is not intelligent. It's a machine. It is not treated like a small child. It is just code performing a task that can be turned on and off.

AI is not a person. It's a machine. If you roll a tire down a hill, you are 100% responsible for what it hits.

In this case the machine created text... While offensive it's not a crime. This is blown away out of proportion.

I'll try to remember this the next time the AI peddlers try to sell me AGI in "just a few years".

A language model is not an intelligent entity.

Seems like no matter what people will continue to make reductive statements like this.

Simple(r) minds need simple world to feel comfortable in it, hence why populist politicians gain so much votes with ridiculous and often stupid claims with success that smarter people just can't comprehend.

Complexities of reality are often frowned upon (yes, its not a very nice comment but then again comment I react on above yours ain't brightest one)