Whoever made the decision was it to implement them as an actor, most likely?

> Whoever made the decision was it to implement them as an actor, most likely?

Directly, the holders of the DAO tokens. That’s what this case is about. Those holders may have further claims against the people who implemented the code or AI. But the liability begins with the general partners, i.e. the holders of the governance tokens.

Will be interesting to see the first judgements enforced against holders of a DAO.

Yeah, booby traps are the first example I can think of illegal automation that comes to mind where the property owner is liable. Setting up something and not pulling the trigger yourself doesn't absolve you.

The people who own it.

Which, in this case, are the bag-holders. Looks like there are benefits to having a real corporation, instead of trying to re-invent corporate governance from scratch, without any input from the courts...

The last people with intent who touched it, usually, which may mean the developers. Certainly that's what happened to the TornadoCash developers.

You can't just wind up a toy with a bomb strapped to it, walk away, and disclaim the consequences; nor can you do that with code.

What if we can't tell who the last person was? Do we just go down the line until we find someone to blame? What if that person didn't add any of the menacing features of the code?

What if the last 'person' was gpt4? is openai responsible?

The person responsible will be whoever the courts think is responsible. Courts use case law and judicial precedent to determine responsibility for a wide variety of complicated legal situations.

Things do get murky when there isn't a lot of existing precedent. That's why it's risky to engage in activities that are, let's just say "legally innovative."

If they can't find the person solely responsible for something, then they got away.

If you make an AI that is hooked up to the internet and it does something criminal, you'd be liable. You can't just let your pitbull roam the streets and not be held liable when it bites someone.

What if I release the AI onto the internet, and then tragically die in a car accident? Who is responsible for the AI? What happens when the AI spawns more AI, who is responsible for that?

Yeah this reminds me of a thought experiment with Ethereum virtual machines. If someone writes and produces the code for tornado cash, publishes it on the internet as free speech but never deploys it obviously they are not at fault. So lets say they publish it and some anonymous person from Somalia publishes it to the Ethereum network and the nodes absorb the program.

When then is responsible for the tornado cash instance? The Ethereum VM node operators who have no idea they're even running it? An unknowable person from Somalia?

Most likely the node operators would be held to be liable if they're hosting illegal content and take no steps to prevent it.

Right now the node operators worldwide even to this day are running tornado cash on their EVM, including validating transactions to sanctioned contract addresses.

And keep in mind, the US Treasury has said that it would issue licenses to Americans who need to be able to withdraw their legitimate funds from Tornado cash. Some validator needs to be able to process those transactions, and would have no idea whether any particular TC withdrawal is licensed or not.

> Who is responsible for the AI?

A deceased person.

> What happens when the AI spawns more AI, who is responsible for that?

The creator of the first copy.

The owners.

For example: Imagine I have a cow, and the cow escapes and goes on a rampage. I didn't make the cow escape, I didn't tell it to go on a rampage, and cows think for themselves so are autonomous.

That doesn't mean I get to claim the cow is liable for the damage.