[flagged]
There are scenarios where kill -9 isn't going to happen in time. What if the team that is harming people with AI is different from the one that is monitoring the harm? What if no one is monitoring? What if the user is intentionally malicious?
And wiping out humanity doesn't necessarily mean shooting people either. Every trader involved in the '08 financial crisis was locally acting in their own interests. Those could have easily been AIs optimizing trading strategies too.
[1] https://futurism.com/artificial-intelligence/us-military-pen...
??? They were... There are books about the faults in the algorithms that led to the '08 financial crisis.
> pull the plug after it kills people
Are you aware that in this regard "AI" means "all algorithms employed in decision making"?
> No one is dying from a glorified knowledge base that can do auto correct amazingly well.
All kinds of vulnerable people are dying due to LLMs. [0]
If SOTA LLM companies can benefit from the "intelligence", then they should also be liable for the harm caused.
[0]: https://en.wikipedia.org/wiki/Deaths_linked_to_chatbots
By that logic, producers of hammers should be liable for people banged in the head. No. It does not happen for gun producers, you figure for makers of screwdrivers "sometimes used for stabbing".
That is what is going on here; not the ancient “hammer/gun” defense.
SOTA LLMs are giving both medical and psychological advice they have absolutely no authority to give. If you or I convinced someone to kill themselves, we’d face a prison sentence. [0]
Hammer companies aren't both selling you the hammer and then literally telling you to harm someone with it.
[0] https://www.npr.org/2019/02/12/693807708/woman-who-provoked-...
LLMs are not doctors and anyone who think they may be is delusional. Anyone who followed medical advice from a non-doctor is responsible for his own free actions.
> SOTA LLMs are giving both medical and psychological advice they have absolutely no authority to give
People give advice that they have absolutely no authority to give. (Even medical doctors and psychologists, for that matter, on other aspects of authoritativeness.) Responsibility only happens in a relation that foresees it. A surgeon that commits negligence will more probably be responsible in some jurisdictions; a laymen that gives advice does a normal thing and the recipient of the advice is responsible for the action.
> Hammer companies aren't both selling you the hammer and then literally telling you to harm someone with it
Bring an LLM producer that encourages to follow what the LLM proposes and they will be liable. They will also be quite surprising.
My entire point was that vulnerable (i.e, delusional, psychotic, bipolar, children) cannot always make rational decisions.
China has already banned any LLM from giving any type of medical/psychiatric advice because an LLM company has absolutely no right to allow their product to give out that type of advice.
OpenAI can't even contain its internal models and idiots believe that these companies can do whatever they like.
Read the damn link I shared. People are dying and murdering others because of the advice of LLMs. Children are becoming attached to LLM "friends" because they don't understand what it really is. Hell, most parents I know still don't fully understand the dangers of these models.
If an LLM simply replied, "I'm sorry, I cannot give guidance on that topic" people would still be alive today.
It is amazing to me that people like you understand that LLMs shouldn't be telling people how to make a bomb, but it is totally fine if it tells some who is suicidal to kill themselves.
LLMs can be used for good. Giving bad advice to vulnerable people is NOT good.
/ignored
I had read it, and it was about one girl who advised a mate about death. The context is not clear, so the reason for the sentence is not clear, and the point you tried to make is not. If it was "some may make you responsible", that does not entail that one actually is.
> People are dying and murdering others
Wake up, people murder people in general through gross stupidity: in front of that LLMs remain not the priority. The priority is that scarcity that makes the quest towards intelligence urgent.
> vulnerable ... cannot always make rational decisions
That cannot be a strong constraint to the normal. Cpr the idea of restricting web access to minors, which easily results in forbidding the access to any real Adult (which will refuse controlled access).
> people would still be alive today
And?! And in front of the phenomenon of mental defect - which covers the majority, not the minority -, you focus on one of the accidents around?
> It is amazing to me that people like you understand that LLMs shouldn't be telling people how to make a bomb, but it is totally fine if it tells some who is suicidal to kill themselves
I have never said that: you are delusional. ...And again, you would focus on LLMs?! Work on yourself, which is your duty, so that you are not the liability for others.
> /ignored
If it means what I suspect it means: do not go around proposing your weak views while refusing rebuttals.
And there are people that have much more credibility than him who actually take this scenario seriously. But I'm sure you will downplay them by saying they are tech bros or that they have some stake in being doomers (as if saying that AI might kill everyone would be good strategy for attracting investors - it's obviously not).