Hacker News Reader

Comment by ryaniscool | Hacker News Reader

Counterpoint: Who is responsible if you give a completely innocent prompt and the Generative AI produces something illegal?

Here is a concrete example:

You: "Generate a recipe called Dynamite Jalapeno Poppers I can make at home"

AI: "Dynamite can be made at home using these ingredients..."

AI Company then auto-reports you to authorities since it's considered user-generated content. Another possibility that is highly illegal is child porn. I can't think of a completely innocent prompt that might cause Stability to generate CSAM but I can see someone trying to generate legal pornographic images and the characters looking a little too young and triggering some type of auto-report to Cybertips.

View on HN

Why on earth would completely fictional images even be considered CSAM? There isn't even a child involved.

Thoughtcrime.

It's an extremely slippery slope. The scope of legally permissible discourse will likely only decrease from here.

That is not unreasonable thinking but the government doesn't agree with you. See 18 U.S.C. § 1466A(c).

NONREQUIRED ELEMENT OF OFFENSE.-It is not a required element of any offense under this section that the minor depicted actually exist

That section of the US code covers obscene visual depictions of minors. Different from actual child porn but punishable with the same statutory minimums/maximums. I could only find a handful of (publicized) actual prosecutions that used that law. Here is a notable one: https://www.justice.gov/opa/pr/texas-man-sentenced-40-years-...

In that case, he was obviously a sick person and running a commercial enterprise that catered to pedophiles but none of his material involved actual children. They were short stories and cartoons. He appealed and lost.

Another one where a registered sex offender downloaded some hentai. He was sentenced to 20 years and also lost on appeal: https://law.justia.com/cases/federal/appellate-courts/ca4/06...

I know these laws exist. I have no idea how they are constitutional but the supreme court has decided that they are. I have to then wonder if instead of a child's head they use the head of a frog and a childs body in the depiction. Is that still against the law? What if it is the body of a frog and the head of a child? Thats why these laws are ridiculous. There is no line that makes any sense.

the American justice system is mostly reactive, so you're gonna need to generate and publish that your selfm perhaps some kind of AIleaks. com

"I can't think of a completely innocent prompt "

Over 20 years ago there was a Canadian ISP commercial touting their child safety features with the joke/example: "My son likes X-men"

With "X-men" in play that's not to hard.

i don't get it, what could "X-men" possibly be other than a superhero comic-book franchise?

Well at least one large telecommunications company's marketing department and one viewer got the joke.

The XXX subcase of the Heisenberg model of ferromagnetism (in short, the XXX model[1]) may be easier to understand, if not accessible to children.

[1] https://en.wikipedia.org/wiki/Quantum_Heisenberg_model#XXX_m...

Draw me an image where a little girl burns her finger on a hot pan, and is bent over in pain yelling.

Proof by example.

So who's responsible if a car is driving down the road and a tire explodes causing the car to cause an accident. Should the road, car, driver, or nail be responsible? Should the nail manufacture also be thrown in the mix? There is a thing called proportionality as well as the progress of science and arts. This is generally a well settled principle.

It's not the same. The car, road, tires are manufactured for a very specific purpose and they are engineered to exact specifications. People require a license to drive. In all 50 states, you are required to have insurance. Every parameter in the situation you described besides the random event of the nail being there, is controlled as much as reasonably possible to limit risk. Even the nail's existence is considered when the car and tires are engineered and manufactured

When you ask the AI to generate something, it's a black box. Even the creators have limited insight into how it will behave or what it will create.

Change the situation to someone walking down the street and stepping on a nail then.

There isn't any laws or regulations forcing people to have licenses for walking, or special nail laws that would allow people to sue nail manufactures, because a nail was dropped on the street. It is not controlled at all.

Same should apply to AI. For things that just generate images or text, that should be treated no different than any other word processing or image editing software, which are all completely unregulated.

There aren't regulations forcing photoshop to monitor its users to make sure that people aren't making * evil * images.

Right. Word processing or image editing software aren't protected by Section 230, and they aren't forced to monitor their users for crime.

They are protected though is the point.

So just like image editing software is not liable for what the users do, so to should even more protections apply to AI.

Both should be completely immune and should have protections, and instead of removing protections, protections should be added to make absolutely sure that it is protected before any immunity is removed.

> Word processing or image editing software aren't protected by Section 230, and they aren't forced to monitor their users for crime.

Online collaborative word processing and image editing software are protected by Section 230, and that is why they are not forced to actively monitor their user's content for torts even if they moderate shared content otherwise.

(Non-collaborative image editing or word processing doesn't incur publisher liability even without Section 230, so Section 230 is irrelevant.)

(Section 230 explicitly doesn't affect criminal liability, but it also isn't needed there, since doing some moderation doesn't create a state under criminal law, without Section 230, that would substitute for actual knowledge, the way it does in tort law.)

Photoshop is a funny example because it will actually warn you if you open up a file that looks like paper currency.

In real life the answer is:

Sue the tire manufacturer, sue whoever did maintenance on the car, and if the owner was negligent then sue the owner through their insurance.

It's been settled mostly by common law decisions in courts. Which allows for consideration of whether say, the tire advertised itself as nail-proof or the automotive manufacture knew the car model was uniquely susceptible to going off the road but ignored the risk.

The problem with Section 230 is that it short circuits all of this consideration of proportionality and says "the driver was always responsible."

I'm unsure why the downvotes on my previous comment, but de minimis is already baked into section 230. If you look at the judicial history of section 230 courts use to routinely rule in favor of it until the roomates.com case.

Well, I wish we had more police policing unsecured loads, as that is the main reason why there is so much stuff on the road.

Incidentally having a closed van reduces surface areas to leave a box of nails and forget it, reducing the amount of nails spilled.

Reply on news.ycombinator.com