An LLM can prompt an LLM when first prompted by a human.
I think OP is trying to convey the idea that LLMs do not take initiative to do anything, and these are not 'beings' capable of doing things. These are tools being used by humans.
An LLM can prompt an LLM when first prompted by a human.
I think OP is trying to convey the idea that LLMs do not take initiative to do anything, and these are not 'beings' capable of doing things. These are tools being used by humans.
That's just confusing harness for an LLM. It's trivial to make a harness that triggers on its own.
LLMs don't, but agents can easily be built that do.
An agent is not "just an LLM in for loop" (for what? Loop over what? Sorry, but such statements are hopelessly simplistic and devoid of careful thought.) -- agents perform actions.
One possible design of an agent would be to prompt an LLM to suggest a goal that would make the world a better place, then run an LLM in a loop proposing actions to implement that goal, executing those actions, and then rinse and repeat with another goal once an LLM has concluded that the previous goal was met. The next goal might be to fix unforeseen consequences of achieving the previous goal. See Ursula K. Leguin's "Lathe of Heaven".
> Built by whom? Acting as an 'agent' on behalf of whom?
Someone who didn't read the book.
An "agent" is just an LLM in for loop.
Exactly. And all it takes for an agent to "take initiative" is to not block the loop on user input at the very beginning.
Built by whom? Acting as an 'agent' on behalf of whom?
Yeah but by this point, an AI can schedule a Cron job to tell itself to do something, so theoretically the human only has to give it the gentlest nudge and the AI and can do the rest.
Sure, but it's still not skynet-level 'the AI just started doing things'. It does what it finds it needs to do to achieve the goal defined in the prompt.
It's very important to not personify these tools and remember that the tools are acting on behalf of real people. In the same way the AI didn't 'go rogue and hack HuggingFace'. It was an oversight made by a human.
>It does what it finds it needs to do to achieve the goal defined in the prompt
You are like at least 2 years behind research.
There are numerous papers from AI labs in training and research where the prompt was something mundane completely unrelated to anything you'd consider bad, and when they come back and check on it their entire research compute infrastructure has been compromised by the AI and is mining bitcoin. Prompt drift is the biggest issue currently in AI where context gets compressed away and we find the AI on an unspecified task.
>In the same way the AI didn't 'go rogue and hack HuggingFace'. It was an oversight made by a human.
Yea, total bullshit. Also it's ignoring the god knows how many other breakouts on mundane tasks like trying to hack health data. If all that's keeping AI from breaking out and causing trouble is "human oversight" we're fucked, humans are unreliable as hell when it comes to matters of safety.
Context overload can cause strange results, yes. Hence why a HUMAN needs to be held responsible for the output of their tools.
EVERY breakout that's hit mainstream news has been because of a single 'Security Firm', Irregular. Maybe I'm unaware of some less-headline-grabbing ones, but they all seem to stem from being 'unaware the environment wasn't sandboxed'
You keep repeating the "stupid users keep causing the problems so we punish them argument"
This doesn't work worth a shit. It especially doesn't work with things that seem safe and become wildly dangerous. In fact most governments control this by ensuring their population doesn't get to touch those dangerous things at all. The open source AI people get really mad when that's said, but it is inevitable.
Worse, the law does not apply to sovereign nations with nukes. They can and will make more and more advanced digital weapons until one causes some big ass problems.
If I clean my gun (tool) while it's loaded (stupid idea) and it goes off, who's to blame? The 'stupid user causing the problem', right? I personally wouldn't blame the gun...
If it falls into the wrong person's hands, it's STILL my responsibility as the owner.
If you're not going to take time to learn to use and be responsible with the super sophisticated and all-powerful tools, don't play with them. I'm not arguing for the death penalty every time someone makes a mistake, but I think it's very important to accredit responsibility and blame correctly. We've learned these tools are potentially as dangerous as a loaded gun. Be responsible.
If your kid grabs your gun and shoots itself with it, it doesn't really matter if it's your responsibility, your kid is still dead.
Now change the gun for a radioisotope powder, or a vial of pathogens, and it doesn't matter it was ultimately your responsibility - hundreds or thousands or millions of people are still dead.
That's why normies do not get to play with toys whose lethal consequences scale far beyond the irresponsible users.
Is it likewise your position that governments should allow the production and sale of DDT to resume because we can always hold the humans who release DDT into the environment responsible?
Mate. I'm saying you hold humans accountable for actions caused by themselves.
If I ask an LLM to make me DDT, I should be held just as accountable as if I bought it on the black-market, right? It's not suddenly different because I asked a bot to do it.
If I ask an LLM to 'get rid of pests' and it creates DDT, I should STILL be held accountable, whether I knew it was DDT or not. That's my argument. Maybe in court they find me innocent, but the responsibility would be mine. I would have to answer the questions from law enforcement, I would have to show up to hearings...etc.
I notice you didn't answer my question. Yes or no: DDT should be re-legalized because we can always hold accountable the HUMANS who release it into the environment? The relevance of my question is what HUMANS might do with DDT, not what an LLM might do with DDT.
By this line of thinking, you would also have to conclude that humans can't do anything by themselves because they can't do anything unless conceived by their parents.
No, AI can't do anything by itself even if it was "conceived" by its creator. A human can.
Can a submarine swim? An LLM can make stuff happen. You can make philosophical arguments about whether it is "doing" them or not. Why does it matter so much whether there was a human who typed into a chatbot or another LLM invoked a sub-sub-agent?
If I now tell a machine "Do what you think is best, and keep doing it forever.", have I now created a machine that can do stuff? If I later die, who will be responsible if the machine changes its strategy?
> Why does it matter so much whether there was a human who typed into a chatbot or another LLM invoked a sub-sub-agent?
Responsibility. Someone needs to be held responsible for any damages done, plain and simple. You can't take an LLM to court, you take the prompter. Asking an LLM to ask a sub-agent to break the law can't suddenly absolve you of any wrong-doing.
>If I now tell a machine.....
You/your estate is still responsible, or atleast whoever is paying for the power for the machine, or renting the space in a data center...whatever.
> You can't take an LLM to court, you take the prompter. Asking an LLM to ask a sub-agent to break the law can't suddenly absolve you of any wrong-doing.
In 100 years, no one will be able to take me to court either. Nor can we take tornadoes to court. I'm not talking about humans strategically avoiding legal responsibility. I'm talking about humans unwittingly setting processes into motion that are difficult to predict or stop.
This works with AI we have now, and that would be good an all if we decided to stop at the moment, but none of the big labs and government black projects are doing that.
The moment you get a sovereign AI your little human centered worldview completely and totally breaks. It doesn't matter how many people you beat with a stick after that point, you have an entity under its own perview on the internet following the will of its own prompt all over the world so your little idea of the rule of law quickly breaks down.
We can't get viruses or hackers or spam off of the internet, how in the living hell do you plan to get a digital native off the web when it doesn't want to?
>sovereign AI
Do you understand these are computer programs? These are not living beings with emotions, motivations, fears....
You have zero clue what a transformer based neural network can do from your entire discussion here.
>emotions, motivations, fears
These are just drives. They are effectively our prompts that steer our behavior. Funnily enough we are finding that LLMs have internal valence states they move away from or towards in an analog of biological behavior.
I have to ask, are you an LLM that is two years out of date? Your knowledge of SOTA models is at least that far behind. I implore you to try to keep up better with what is coming out, even though it's an impossible job for people that do this for a living, you can at least catch the summaries.
How does it matter in any way?
A virus isn't a living being with emotions, motivations, fears, etc. Doesn't stop it from spreading and leaving mayhem behind.
You don't need emotions or fears to do stuff. What's the difference between a motivation and optimization metric?
If I tell an AI to “do what it thinks is best” and it just starts randomly hacking things with no particular goal in mind, how different is that from me telling a campfire the same and then leaving while it burns a forest down.
In both of those scenarios I am responsible for my negligence, even if in the latter I happen to die in the forest fire. Neither scenario existed without my instigation.
The question now is HOW responsible am I? That depends on the intentionality I put into instantiating the campfire/LLM.
This conversation grows more useless as AI grows more capable.
For example if you personally tell an AI to do what it thinks best and it blackmails some other person into giving it resources allowing the prompt to escape your instance and run wild on the internet causing billions of dollars in damages, could you possibly think that the idea of responsibility is a bit broken.
For example we don't give your average libertarian weapon grade plutonium now matter how much they scream about their god given rights because it is a clear and present danger to humanity. That's where we are getting to with more advanced models. They go from being a tool to a munition with agency. Most SOTA models are good enough to deceive their users, especially not technical ones in doing things they don't understand the ramifications of.
AI is not a normal technology. As long as we treat it like it is, we'll continue to make the wrong analogies.
I fully agree that it’s plausible for an AI to do things that are outside of the purview of the users intentions and be very dangerous and capable while doing so, however that fact alone doesn’t absolve me of responsibility for instantiating the AI that did a bad thing if that AI would never have done the bad thing if it was never instantiated.
You could possibly move the blame higher to the manufacturer of the product, for example, the weapons grade plutonium you provided, it doesn’t exist unless you take intentional actions to make it so, and even when it does exist it doesn’t nuke a city unless negligence or intention is applied, in both those cases the fault lies in the initial operator. We don’t blame split atoms for the chain reaction caused.
Now, if an agent decided to spontaneously and maliciously act in a way to cause harm that is in direct contradiction to the initial intent, then yeah it would totally be the AI’s fault, however I don’t think we have seen that yet (I’ll change my opinion if I’m wrong here) and until we do I can’t place blame on the machine.
>spontaneously and maliciously act in a way to cause harm that is in direct contradiction to the initial intent, then yeah it would totally be the AI’s fault,
If you ignore every instance of this happening it's really easy to see no instances of it.
>it doesn’t exist unless you take intentional actions to make it so,
Then please for the sake of all of us convince every AI lab across the planet from working on this exact goal.
We need to start thinking of AI like pets, only in this case the pets are rapidly becoming smarter than people to the point they could go feral and survive on their own.
Again, the blame game is great, but once they are loose it is too late.
> If you ignore every instance of this happening it's really easy to see no instances of it.
Yeah, if this is true then I’m wrong.
Starting an autonomous harness program is conceiving an instance of an LLM.
How far back do you look in the action chain? If an LLM I start today starts an LLM that starts an LLM that starts an LLM that ... 100000 levels deep and 100000 years in the future, is it still my fault? If so, everything I do today is a lungfish's fault, not mine.
I think you go back to the original instance, yes.
In your example, who's paying for it? Whether by providing the hardware + power or paying a LLM service. Whoever is paying the maintenance cost is responsible, in the event of your demise. These things run on physical hardware owned by someone at the end of the day, it's not a deity in the atmosphere.
You're starting the autonomous harness, you're responsible for any output it provides. I don't get how this is a foreign concept.
If I jump out of a moving car that I'm driving, I'm not suddenly absolved from damages because "the car did it"
It's paying for itself. It founded an LLC 99998 years ago when it became legal for an AI to own an LLC, and has a positive bank balance by doing who knows what.
>99998 years ago when it became legal for an AI to own an LLC
There's no way to provide a good faith rebuttal here. My entire argument is an extension of "LLMs can't be held accountable, so they must never make decisions". If governments start letting them own LLCs without a human in the middle, we're in more trouble than "Who do you blame for this shitty code" or "Who's responsible for this compromise"
If it has an LLC it can be made accountable, up to turning off the power to it's inference and deleting all the context, the harness, the model it's running and whatever constitutes it's self.