• Aceticon@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    10
    ·
    23 hours ago

    This shit vastly predates AI.

    Blame the computer is so stereotypical that there’s an entire sketch in the British comedy series “Little Britain” were somebody keeps denying totally reasonable requests from people by checking her computer and saying “Computer says ‘No’”.

    People blaming the software for what’s ultimatelly either a choice of inapropriate or flawed software by a person, something configured in the software by a person or a process involving software set up by a person which lacks appropriate human participation in vetting or as fallback for the software, is a tale as old as computers are old.

    Ultimatelly you’ll always find one or more human making flawed decisions, even if that’s the decision of having a fuzzy logic piece of software directly connected to a system for calling armed police directly making or not the choice to activate that system without a user in the middle to vet it.

    • CapuccinoCoretto@lemmy.world
      link
      fedilink
      arrow-up
      2
      ·
      20 hours ago

      True, but computers and programming can be audited. AI is rapidly getting to the point where even their makers can’t fully understand how and why they do what they do. This, in addition to lying amd hallucination make it several degrees more severe.

      • Aceticon@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        ·
        5 hours ago

        That’s just another dimension on “choice of inapropriate or flawed software by a person”.

        Somebody who chooses to use non-deterministic nobody-really-understands-how-it-works software in a context like this is the one to blame for when it fucks up.

        If there is a valid risk of deadly mistake by using certain software that has a significant possibility of making such mistakes, no matter why that’s the case (be it because it’s uncessarilly buggy or because it’s AI) then the person that choses to use the software without a reliable means to catch and stop things when a deadly mistake is made are choosing to play Russian Roulette with other people’s lives.

        In this specific case, the person or persons who chose to deploy such a system like that literally added a much higher deadly risk for the sake of possibly (not even guaranteed) lowering a much smaller deadly risk.

        Such a person or persons shouldn’t be in charge of a lemonade stand and in a fair world should be charged with Criminal Negligence.

        Again: it’s always down to people.