Hi! To find out what I can do, say @discobot display help.
why?
the (legal) system took competently care of this nonsense … see, its already baked into our civil-law-codes … no need for ever more texts in the legal system.
I am surprised that another “civil correctional system” (social-media) did not get triggered - as normally those shitstorms are fairly swift correctional-events that companies have a keen eye on.
I am really surprised this matter blew all up in a legal process (iow: no Air-C mmgt did stop it from getting there by refunding - say $500,- )
That precisely described AT&T prior to its breakup in 1984. Our lines, our phones, STFU and conform, or else.
Now, can we get a sensible legal ruling on the US notion that corporate entities are ‘persons’ under the law?
This claim by AirCanada is ludicrous on multiple levels. Suppose that it were a separate entity. Suppose, for instance, that instead of talking to a chatbot, this guy had talked to an actual human being for customer service. If that actual human being, with a separate legal existence, had told him the wrong fare information, then AirCanada would still be liable for what the human told him. If that human told him wrong information, they could fire the human, or even sue the human for the amount that the inaccurate information cost them, but they’d still have to honor it.
The Tort system is all you need to solve this, but it’s going to take some more legal battles before we sort out the chains of responsibility when AI is in the mix.
Consider a situation where OpenAI builds a foundation model. Then someone uses that model and fine-tunes it for their business case, and puts it to work. The new AI hallucinates and kills someone in a factory with a bad robot command. Who is responsible?
A) The company, for putting AIs in a safety critical job, when they should know better.
B) The people who fine-tuned the model
C) OpenAI, for releasing a model capable of hallucinating without offering suitable safeguards or warnings.
When I first heard that they tried to say the AI was its own ‘legal entity’, I thought they meant that they are using a fine-tuned OpenAI GPT, and therefore thought they aren’t responsible for the bad information it put out. But if they were running their own LLM on their own hardware, that’s ridiculous.
The legal system is going to take a long time to catch up to AIs, I think. But this case seems much simpler. It doesn’t matter if the AI is or isn’t its own legal entity, because once it was put on Air Canada’s website as an Air Canada advisor, they have to assume responsibility. Just like they would if it wass a human contractor from another company in their call center. Now, if it was an OpenAI GPT and Air Canada thinks its ability was misrepresented, they could then try to sue OpenAI themselves. But that doesn’t get them off the hook for their legal liability to the customer.
I imagine the line of thinking never got this far, since the defense asserted was risible on its face, but even if someone went batshit and accepted the theory that the AI was its own legal entity, it still doesn’t absolve AC. It’s like their lawyers never heard of agency.
Air Canada has been found liable for the chatbot’s mistake.
Or rather they hoped the customer had not. But of course the court had.
I encountered the chatbot from my cable company once.
I ended up driving to the service office. Even counting drive and wait time I could have solved the issue in half the time if I had done that to begin with.
If the chatbot is responsible for its own actions, what happens when robots start flying the plane?
They start annoying everyone at parties by telling them they are a pilot over and over again?
I knew one who would literally interrupt any conversation where a place name was mentioned so he could say “I could fly there”.
Sheesh, what pathetic losers. I never volunteer(ed) my job until / unless asked. Once they know, folks love to ask me things about it a lot more than I like(d) to talk about it.
As the old joke goes:
Q: “How do you know if there is a pilot at your party?”
A: “Wait a minute. They’ll tell you.”
I’m sure there are many people this does not apply to. I suspect it is more amateurs than the professionals.