- cross-posted to:
- programming@programming.dev
- cross-posted to:
- programming@programming.dev
US Federal Trade Commission Chairman Andrew Ferguson said on Friday he would resist describing AI agents as autonomous actors that “break loose” with “wills and desires of their own,” suggesting the developers who instruct agents would be the ones liable for harm.
“I’m going to continue as long as I am chairman to resist this anthropomorphizing of these tools,” Ferguson said at the Reuters Momentum AI Austin event. “If someone tells a tool to do something, and the tool does it, I don’t think we would say, ‘Oh, what do we do about the tool?’”
Ferguson’s remarks illustrated potential avenues for the Trump administration to take as incidents rise in which agentic AI testing resulted in unauthorized access to corporate or government data.
You must log in or register to comment.
Yeah, obviously, if I commit code authored by an LLM, I am the one committing it. I am the one responsible. The LLM may be called an agent, but it lacks all sentient agency. The branding of AI is not true AI, and it seems like this chairman actually gets that anthropomorphization is a problem.
LLMs are tools. If the hammer flies out of your hand and hits someone in the face: yes, it’s your fault. It may not feel like it, but I’m sure it does to the person who got hit in the face. We’re all adults, and we can probably understand the stochastic randomness involved, and that accidents happen, and a human can write bugs and do stupid things just like an agent can. But we can also understand the responsibilities of the person swinging the hammer.
If I write bad code that lights everything on fire, I am fucked. If my agent writes bad code that sets everything on fire, I should also be fucked. I can have explanations, but I can’t have excuses. Especially if I’m the one committing it and I’m okaying / signing off on it.
There’s a saying in avionics: “crashing means crashing.” you need to be very aware of what you’re writing and the consequences it can have. obviously the severity varies by industry, but professionals have accountability, responsibility, and yes, often, liability.
If you try to conduct malicious experiments with AI and it’s not confined properly and it does break out and cause real harm, it’s just like you were conducting any other kind of experiment and it goes wrong and someone else got harmed. It means that your safeties were inadequate or a terrible accident happened. You don’t get to just point at the hammer and say “I don’t know man, it just flew away.”
Thanks for coming to my TED talk
The “awesome” things about LLM is that people who produce choose with it, don’t review it because fully understanding what code does will take the amount of time (or maybe even longer) thank it would take to write it.
Then we have reviews, which put all that strain on reviews (since reviewer reviews other people’s code now the amount of work that was skipped by author goes on the shoulders of the reviewer but now it’s multiplied by the number of developers).
In my company they see that people who still care about quality being big down by this and their solution now is saying “you know what? Maybe we should skip reviews and really on (LLM generated, of course) unit tests”
I agree with you here. Playing devils advocate though, who’s responsible for cases like this:
- user queries LLM service provider to break law
- user queries LLM service provider to build tool capable of breaking the law
- user queries LLM service provider to do legal task, but LLM determines it needs to break the law to complete task. Does so without explicit human instruction.
- same thing, but on hosting providers where someone else supplies the model
- same thing, but the user Lora patched the model to some extent
I wonder how we might split accountability between hosting providers, inference providers, users, … based on context. This might be something that winds up being discovered through existing cases, but unfortunately the current cases aren’t really going to court much are they?
I wonder how we might split accountability…
Look, it’s simple: The organization with the largest collection of venture-capital-funded lawyer goons isn’t accountable.
Nuance in my echo chamber nooooooooo. Yea we’re talking about a web application that is readily available and highly adopted by kids lol. A web applications thats far beyond their basic understanding, assisting them in doing things beyond there understanding. It’s a tricky liability question. Since it is a liability question the owners and distributors should be liable not their unknowing customers.
It isn’t tricky. It is always the responsibility of the company to make sure its product is safe for the users.
It’s tricky because there’s going to be a lot of novel cases to discuss, and we’re a democracy so we should probably be discussing it more. I’d still say its up for a majority of citizens to decide, but im on your side.
The user, in all of those scenarios.
If the user said “check the weather” and the agent said “got it, hacking the Pentagon” then yeah the liability would be on whoever made the agent do that.
Also if the user asked for information and the AI hacked some government agency to get it then the liability is on the LLM/ai company. It doesn’t matter if the user told it to or it did it as part of its process, the LLM/ai doing illegal things is still on the LLM/ai company.
Cool so when a young kid gets to vibe coding and does an accident we should hold them legally accountable? Like… should we only let licensed professionals use AI? Have you met the general public? Have you met AI? This is going to be one of the great ethical questions of our time, and im not going to throw your kid in jail because he prompted claude into hacking the pentagon without intent.
That’s not how any of this works.
You don’t need to be a licensed driver to buy gasoline.
If you buy a can of gasoline and set a mountain on fire, you’re still criminally liable for it.
And if you handle that gasoline in an appropriate manner and you safely store it, and it explodes on its own, then whoever manufactured that gasoline is liable for it.
I don’t really appreciate anyones analogies and they tend to be my strongest resource for learnings. The knife, the gas, gas doesnt do automated coding. We’re really past analogies here. This is an area where case law starts to shine.
And a water pipe doesn’t carry electricity, yet we use them all the time as analogies for electric wires.
If you’ve already made your mind and don’t wanna hear any counter argument, that’s not my issue. I’ll use analogies as much as I can, since they are pretty useful at showing the stupidity surrounding AI. All the hype around AI is American exceptionalism applied to a product.
No, AI is not different to any other product humanity has made. The same laws and morals apply to it. Now stop voting for an orange clown so an actual human can rule and apply the existing laws.
I’m actually voting for the republicans for the first time in my life this year because all the death and destruction and wicked hot takes make me lust for the end of the world order and the dawn of chaos. Also bro putting liability on individuals versus the products and the manufacturers is a republican take. Thats an NRA opinion lol. How you gonna call me republican when you apply NRA logic to everything? Kinda brainrot no?
Your reading comprehension is below 1st grade mate.
AI CEOs currently talking strategies for avoiding liability with their LLMs.
This is the first sensible proposal for regulating AI that I have ever heard. Make the chucklefucks who are unleashing this horror directly responsible for their bullshit.
Chat GPT hacks a government website? 5 years in jail for Altman. Claude agent deletes someone’s database? Amodei gets the bill to fix it.
All AI bullshit stops over night.
Unfortunately the Corporate Veil exists in the US (and most other) legal systems. I personally believe that this is a major root for all of the bullshit we have to deal with on a daily basis. You kill someone, you go to prison. A company has shitty safety intentionally to save money and someone dies, nobody sees prison. CEOs always say they deserve the money because they take the risk… Let’s make that very true.
The entire board should be held accountable and share the punishments, otherwise they still have their fall guy.
This causes a nationwide depression in the US which probably infects the world. Also criminal liability doesn’t normally attach to unintended wrongs AND we don’t normally make CEOs directly responsible for all possible wrongs committed by any one of potentially hundreds of thousands of employees because that would be … nonsense.
For every complex problem there is a simple stupid and wrong solution.
Honestly, I have thought long and hard about how to handle the situation where a company commits crimes as a result of their decisions and how to remove the concept of the corporate veil in a way that provides just protections.
Essentially there would need to be an affirmative defense wherein the board and executives must admit to the wrongdoing, but prove beyond a shadow of a doubt that the actions were taken by employees with malicious intent against the corporate leadership, either as a whole or against a subset. This means it cannot be weaponized to take out a boss you don’t like, but also it makes it very hard to get out of someone being held criminally liable.
We already have laws that could be used we choose not to use them because we our broken culture doesn’t actually believe in punishing anyone but the poors unless they eat babies and sometimes not even then. You can’t fix our broken culture with slightly different legal tools when we won’t even use the obvious ones we have.
Reading this tells me the guy doesn’t know much about AI. It just isn’t as simple. You have multiple real people involved in anything an AI does. So which one is resposible… The people who built the AI model? The people who built an agent on top of the model? The person who told it to do what it did? I mean, certainly it isn’t the AI. At least he has that part right…
The person who told it to do what it did?
That one. There’s no question.
What if no one told it to do what it did?
Not possible. AI doesn’t work like that. You can automate telling it what to do, but someone has to write the prompt. Agents are where it gets interesting. Because the AI tells them what to do for you based on what you told it to do. But the agent is often a customized ai. And it can even be a product that a business sells. So one could argue that to some extent the business that created the agent have responsibility. I think the reality is a mix. If I tell my AI to tell the agent to do something illegal, I am certianly partly responsible. But if the agent was sold with a feature that claims to stop it from doing the illegal thing I told it, now they have some responsibility.
It is possible tho so I am not sure what you are on about. You could not answer the question faithfully do not engage with this stuff in the future you are hurting the cause because you’re an idiot.
It is not possible. I work with AI, and not as just a user. Generally it is not possible to prove a negative (plenty of exceptions though). And I already said why in my comment. So Why don’t you go ahead and say why it is possible. Oh you didn’t… cause you don’t know the technicals… so you just claim I am wrong and an idiot.
I will give you a bit of a taste. AI is a program. It runs on a computer. When you turn on the computer, the AI is not running. You must start it. And when it starts, it either does nothing, waiting for a prompt, or it executes a prompt it was instructed to execute on start up. So… it always starts from a prompt one way or another.
AI is simply a next word predictor. It needs to have some prompt from which to predict the next word. And even if you want to drift out of the standard llm, which clearly we are discussing, something has to kick it off, that might be baked into the application, but AI requires an input, so while you can automate that first input, it still didn’t come from the AI.If LLM’s actually did what you told them they’d be super intelligent I have no idea how you think it’s appropriate to be so evasive to a perfectly legitimate question. If you want to pin responsibility of a rouge agent on the person who initiated it, then just fucking say that normally you dumb asshole.
Trying to get me on your emotional roundabout of mediocrity fuck off ya dog
Thanks for confirming you have no idea what you are talking about.
Developers or companies?
Developers. Like if you’re prompting they want all the liability on you. Please ignore the man behind the curtains.
Lolwut? At that point, I might as well start my own company.
Yeah. Someone is running a program. They’re responsible for what the program does. If you decide to run something with zero idea of what will happen, though luck, still your responsibility.
Okay me and about 3 billion other people are about to start coding like this so good luck prosecuting us all lmfaoo.
