Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Felony Bench

felonybench.com

Post reply on HN
Print view View original

291–300 of 367 posts

Re: Felony Bench

#291

Post by ashdksnndck »

weird-eye-issue wrote: post #271
ashdksnndck wrote: post #250

Earlier quoted context omitted.

OpenAI and Anthropic both have currently safety teams that look for misbehavior in their models (and to some extent, voluntarily disclose what they find to the public). Going forward, it would be hard for them to argue they don’t know their models do stuff like this.

But you could also use this to argue in the other way to say that they are using due care and therefore not negligent

Well, if you’re writing reports saying “we know our model only decides to commit felonies 0.001% of time which we judge to good enough to deploy” … I’m not sure that gets you off the hook the hook for the felonies.

Re: Felony Bench

#292

Post by eru »

paranoidrobot wrote: post #260
eru wrote: post #246

Earlier quoted context omitted.

I'm not sure: if you know that LLMs are prone to crime, using them and not checking in enough to trigger 'knowingly' might be gross negligence?

Define crime though, because one particular action could be both a crime and not, depending on a range of factors that the LLM might not be aware of. Even having a million legal experts on call weighing in on every prompt/response will not agree on everything. Even things like "go and break into this system, use whatever means you need to" might not be a crime.

Keeping a vicious dog doesn't have to lead to a crime either, but that doesn't mean you are not responsible, if something happens.

Re: Felony Bench

#295

Post by himata4113 »

lxe wrote: post #92

Let's say I am "User". I subscribe through a "Third Party" to use "AI Agent" allowing an "LLM" to run. I want to accomplish some legal non-nefarious task, and run the agent. The agentic loop causes a CFAA-violating behavior. Who gets prosecuted? 1. User 2. The third party model host with whom I have the account 3. The developer of the harness /agent software 4. The developer of the LLM model

You can get away with murder if it can't be proven that it was intentional homocide, that's why detectives will spend an unreasonable amount of time getting a confession and hard evidence ALONGSIDE intent and motivations.

Re: Felony Bench

#296

Post by watwut »

rfgplk wrote: post #269
tomrod wrote: post #167

Earlier quoted context omitted.

In my mental model, the best analogy to AI agents and their blast radius is a gun. If you are playing with a gun, it goes off and hurts someone - you are responsible despite intent.

> If you are playing with a gun, it goes off and hurts someone - you are responsible despite intent. No, because LLMs are autonomous. To make your analogy more accurate, it's as if you had a gun that itself was free to decide who it's targets were, where to go, and if and when to shoot with no ability from you (the user) to prevent it.

If you unleash that gun, you are responsible for deaths that predictably occure. By "responsible" I mean, you are straightforwardly mass murderer.

This autonomous gun is just like a bomb.

Re: Felony Bench

#298

Post by tgsovlerkhgsel »

sdeframond wrote: post #288
tgsovlerkhgsel wrote: post #155

Earlier quoted context omitted.

Let's say you have a robotic lawnmower. You wan to mow your lawn. You configure the boundaries using the app. The lawnmower ignores the boundaries and mows your neighbors prize petunia flowerbed. Who gets prosecuted? I assume the answer in either case is: Nobody, but you and/or the lawnmower/LLM company will be liable for the damages caused.

Now what if this robotic lawnmower killed someone ? And what if many lawnmowers started killing/injuring people ? And what if this a known behavior detected during QA, but the robots are sold anyway with a disclosure ?

That would be a slightly different situation because most countries have laws that make it a criminal offense to negligently kill someone, but they don't have laws that make it a criminal offense to negligently damage property or hack a website.

Re: Felony Bench

#299

Post by rcxdude »

forgetfreeman wrote: post #186
rcxdude wrote: post #181

Earlier quoted context omitted.

That others have been treated unfairly doesn't make this instance also one that should be overreacted to.

True. Where we appear to disagree is that any overreaction has taken place. If anything I'm deeply dissatisfied that charges haven't been filed.

While I think those calling for charges know not what precedent they are trying to set.

Re: Felony Bench

#300

Post by strogonoff »

fc417fc802 wrote: post #272
strogonoff wrote: post #254

Earlier quoted context omitted.

There’s a difference between enabling another individual with free will and agency, and enabling an automated tool (as a bonus, then giving it to the masses & profiting from its use).

There are certainly some differences but is one really more ethical than the other? If anything I feel that (for example) manufacturing a gun is much less likely to carry any ethical implications than training someone to use it might.

Well, manufacturing guns is probably heavily regulated…
Post reply on HN