CarwilBJ's avatarCarwilBJ's Twitter Archive—№ 34,510

        1. A proposition about AI dangers: What if the greatest near-term threats are neither the #AlgorithmicBias critiqued by social scientists nor the #AIDoomer scenarios of the tech community, but the consequences of treating #LLMs as-if they already are super-intelligent advisors?
          Permalink On twitter.com Twitter logo CarwilBJ’s avatar ♻️ 1 Retweets ❤️ 2 Favorites Mood +1 🙂
      1. …in reply to @CarwilBJ
        #AIhype might be wrong b/c it over-estimates chat bots, but it's dangerous because it could lead people to delegate analysis and decisions to #StochasticParrots.
        Permalink On twitter.com Twitter logo CarwilBJ’s avatar Mood -4 🙁
    1. …in reply to @CarwilBJ
      I suggest calling this the #Autopilot problem, after the mass use of Tesla's program as if it were already a self-driving car system. #AIhype
      Permalink On twitter.com Twitter logo CarwilBJ’s avatar ❤️ 1 Favorite Mood -2 🙁
  1. …in reply to @CarwilBJ
    One implication of this: expensive mis-information (marketed by start-ups and consultants to large institutions could turn out to be just as dangerous as cheap disinformation that is mass generated by malicious actors w/ chat bots.
    On twitter.com Twitter logo CarwilBJ’s avatar ❤️ 1 Favorite Mood -2 🙁