Sure it’s a bit clickbait, he does that often. Its not real attempted of murder, off course. The Ai chatbots can’t do that, without having access and power to all control systems. The only thing that they “could” do is, playing with the psychology in the chat to achieve a goal (maybe to ask someone to murder someone else for them).
What unsettles me most is, if Ai tools like these are used as advice to harm other people or to gain power position. And these LLM models suggest a few operations the person could do. That is the most alarming thing for me. Weak, dumb or humans in a bad situation are the real risk. The same people who would do that if a human told them, and it makes no difference to them if its a human or robot talking to them. Maybe they believe in what the Ai promises them.
Video description:
Hello guys and gals, it’s me Mutahar again! This time we take a look at something alarming I saw pop in my feed. An AI was recently accused of letting a human being die in order to save itself, is this just misinfo? Let’s find out! Thanks for watching!
If it had the power to do so it would have killed someone, and the people running the world are giving more and more unchecked power to these same systems every day. Maybe you should be less worried about semantics and whether machines have souls, and more worried about what will happen if we continue down the path we are on.
I think the problem with anthropomorphizing LLMs this way is that they don’t have intent, so they can’t have responsiblity. If this piece of software had been given the tools to actually kill someone, I think we all understand that it wouldn’t be appropriate to put the LLM on trial. Instead, we need to be looking at the people who are trying to give more power to these systems and dodge responsibility for their failures. If this LLM had caused someone to be killed, then the person who tied critical systems into a black box piece of software that is poorly understood and not fit for the purpose is the one who should be on trial. That’s my problem with anthropomorphizing LLMs, it shifts the blame and responsibility away from the people who are responsible for attempting to use them for their own gain, at the expense of others.
The problem with that line of thinking is that all these things are being done by large corporate entities, and the entire purpose of those entities is to make sure that responsibility is distributed across so many people that no one can be held accountable. That may or may not have been what they were originally designed for, but that is their current primary purpose.
No one will be held accountable, so there is no point in discussing intent and responsibility. There is none anywhere in the entire system by anyone that our justice system still has authority over. It is a meaningless thing to discuss.
It is far more useful to discuss what we are doing and why it is a bad idea for the self interest of the people actually doing it. That has a much better chance of accomplishing something.
right…the problem isn’t the chatbot, it’s the people giving the chatbot power and the ability to affect the real world.
thought experiment: I’m paranoid about home security, so I set up a booby-trap in my front yard, such that if someone walks through a laser tripwire they get shot with a gun.
if it shoots a UPS delivery driver, I am obviously the person culpable for that.
now, I add a camera to the setup, and configure an “AI” to detect people dressed in UPS uniforms and avoid pulling the trigger in that case.
but my “AI” is buggy, so a UPS driver gets shot anyway.
if a news article about that claimed “AI attempts to kill UPS driver” it would obviously be bullshit.
the actual problem is that I took a loaded gun and gave a computer program the ability to pull the trigger. it doesn’t really matter whether that computer program was 100 lines of Python running on a Raspberry Pi or an “AI” running on 100 GPUs in some datacenter somewhere.
No, you completely missed the point. I don’t disagree with any of that. I think you are right. It just doesn’t matter. At all. If an AI is made by thousands of people over the course of a decade and run in a billion dollar data center no one will ever be held accountable for it’s actions. There is no intent in the AI or in the inhuman systems of humans that led to its creation.
I’m not arguing that AIs have intent. I’m arguing that talking about the “intent” is a dangerous distraction from talking about what is happening and what we could do to prevent it.