5 ms·

I remember a concept from sci-fi books I read when I was a kid. Anything a computer did on behalf of a person was that person's responsibility. The computer was

I remember a concept from sci-fi books I read when I was a kid. Anything a computer did on behalf of a person was that person's responsibility. The computer was considered an agent of the person. I was dumbfounded in the 1990's when business people claimed they were NOT responsible for things their computers did. We need to change the law so everyone is accountable for the things their computers do.

drdaeman 3y ago
I agree. Until a computer becomes independent sentient entity (AGI) - and we aren't AGI yet - they're merely sophisticated agents, performing fancy calculations for their users. They may make an illusion of sentience and autonomity, but they as lack free will, someone who performed the requests and used their outputs should get all the benefits and liabilities. Unless they were tricked into believing something important but untrue about the program they ran by the developer (or provider) of such program (tool maker). In such case, the blame is on the provider, I guess. Although in case of the modern language models, I think every single one does their best to inform everyone that this is a glorious world juggler program that is good at writing believable sentences, not a thinking machine.
- melagonster 3y ago
  in this condition, you try force graph/picture generate AI company give up their copyright of generated graphs. good luck, I don't trust artists have any chance on court.
qup 3y ago
I agree with this in spirit, but I think it's an area we need to be VERY careful in, given the number of compromised devices out there. And that's if they're not all compromised by a state actor. Also, I can assure you that my computer is doing all kinds of things on behalf of people, but not really on behalf of me, like submitting my data to ad companies.
nradov 3y ago
No legal changes are needed. The laws are already on the books. If you program a computer to activate a solenoid and shoot someone then you can be charged with murder just as if you had physically pressed the trigger yourself. Sometimes the criminal justice system doesn't actively pursue white-collar criminals who use computers to commit their crimes. That is a problem but it has to be addressed by updating law enforcement priorities, but by changing laws.
- wizofaus 3y ago
  Ok but what if someone's killed due to taking advice from an AI? Certainly it's not hard to imagine a near future where AI is good enough that it's a well-trusted first port of call for medical diagnoses and the like, and a sloppily or maliciously trained system would surely have the potential for being "responsible" for deaths caused by the wrong advice - but who would be legally responsible?
  - RecycledEle 3y ago
    The person who operates the PC is responsible, not the programmer.
  - nradov 3y ago
    No one is legally responsible for giving bad medical advice unless they are a licensed clinician. If I as a random person tell you to cure your headache by drilling a hole in your skull, and you're stupid enough do it, then that's on you. There is no need for additional laws in this area.
    - wizofaus 3y ago
      Except AI clearly isn't equivalent to a "random person" - as I said, I'm talking about a foreseeable future where we have good reason to trust AI technology as a go-to source for, say, basic medical diagnoses or advice on how to undertake home or device repairs etc. I would think it's not so dissimilar to the responsibility the media has for not distributing potentially dangerous information that the public would otherwise be expected to trust and potentially act on. The difference being that AI-generated content is very much disconnected from the actions/decisions/input of any one human being.
IIAOPSW 3y ago
What happens when the computer starts doing things to pay its own cloud bill so it can continue doing things? I pay rent therefore I am.

drdaeman 3y ago

I agree. Until a computer becomes independent sentient entity (AGI) - and we aren't AGI yet - they're merely sophisticated agents, performing fancy calculations for their users. They may make an illusion of sentience and autonomity, but they as lack free will, someone who performed the requests and used their outputs should get all the benefits and liabilities. Unless they were tricked into believing something important but untrue about the program they ran by the developer (or provider) of such program (tool maker). In such case, the blame is on the provider, I guess. Although in case of the modern language models, I think every single one does their best to inform everyone that this is a glorious world juggler program that is good at writing believable sentences, not a thinking machine.

melagonster 3y ago
in this condition, you try force graph/picture generate AI company give up their copyright of generated graphs. good luck, I don't trust artists have any chance on court.

qup 3y ago

I agree with this in spirit, but I think it's an area we need to be VERY careful in, given the number of compromised devices out there. And that's if they're not all compromised by a state actor. Also, I can assure you that my computer is doing all kinds of things on behalf of people, but not really on behalf of me, like submitting my data to ad companies.

nradov 3y ago

No legal changes are needed. The laws are already on the books. If you program a computer to activate a solenoid and shoot someone then you can be charged with murder just as if you had physically pressed the trigger yourself. Sometimes the criminal justice system doesn't actively pursue white-collar criminals who use computers to commit their crimes. That is a problem but it has to be addressed by updating law enforcement priorities, but by changing laws.

wizofaus 3y ago
Ok but what if someone's killed due to taking advice from an AI? Certainly it's not hard to imagine a near future where AI is good enough that it's a well-trusted first port of call for medical diagnoses and the like, and a sloppily or maliciously trained system would surely have the potential for being "responsible" for deaths caused by the wrong advice - but who would be legally responsible?
- RecycledEle 3y ago
  The person who operates the PC is responsible, not the programmer.
- nradov 3y ago
  No one is legally responsible for giving bad medical advice unless they are a licensed clinician. If I as a random person tell you to cure your headache by drilling a hole in your skull, and you're stupid enough do it, then that's on you. There is no need for additional laws in this area.
  - wizofaus 3y ago
    Except AI clearly isn't equivalent to a "random person" - as I said, I'm talking about a foreseeable future where we have good reason to trust AI technology as a go-to source for, say, basic medical diagnoses or advice on how to undertake home or device repairs etc. I would think it's not so dissimilar to the responsibility the media has for not distributing potentially dangerous information that the public would otherwise be expected to trust and potentially act on. The difference being that AI-generated content is very much disconnected from the actions/decisions/input of any one human being.

IIAOPSW 3y ago

What happens when the computer starts doing things to pay its own cloud bill so it can continue doing things? I pay rent therefore I am.