Here’s a new preprint from Anthropic: “How AI Impacts Skill Formation”. AI coding bots make you bad at learning, and don’t even speed you up. [arXiv] The researchers ran 50 test subjects through fi…
    • fiat_lux@lemmy.worldEnglish
      171·
      1 day ago

      I read both of these and what struck me was how both studies felt remarkably naive. I found myself thinking: “there’s no way the authors have any background in the humanities”. Turns out there’s 2 authors, lo and behold, both with computer science degrees. This might explain why it feels like they’re somehow incredulous at the results - they’ve approached the problem as evaluating a system’s fitness in a vacuum.

      But it’s not a system in a vacuum. It’s a vacuum that has sucked up our social system, sold to bolster the social standing of the heads of a social construct.

      Had they looked at the context of how AI has been marketed, as an authoritative productivity booster, they might have had some idea why both disempowerment and reduced mastery could be occurring: The participants were told to work fast and consult the AI. What a shock that people took the responses seriously and didn’t have time to learn!

      I’d ask why Anthropic had computer scientists conducting sociological research, but I assume this part of output has just been published to assuage criticism of trust and safety practices. The final result will probably be adding another line of ‘if query includes medical term then print “always ask a doctor first”’ to the system prompt.

      This constant vacillation between “it’s a revolution and changes our entire reality!” and “you can’t trust it and you need to do your own research” from the AI companies is fucking tiresome. You can’t have it both ways.

        • scruiser@awful.systemsEnglish
          4·
          10 hours ago

          “How AI Impacts Skill Formation” has two authors. So even on the bare factual matters you are wrong. The disempowerment paper has four authors, but all of them look like they are computer scientists from looking at their bios, so the general thrust of fiat_lux’s comment is also true about that paper.

          I don’t mind academics reaching outside their fields of expertise, but they really should get collaborators with the appropriate background, and the fact that anthropic hasn’t hired any humanities researchers to help support this sort of research is a bad sign.

        • blakestacey@awful.systemsEnglish
          7·
          13 hours ago

          Awful.systems is not debate club. Nor is it peer-review club. No one is obligated to nitpick individual sentences in a preprint or erect monuments of text about details within it, particularly when a discussion of the broader failings of the “research” culture in that area is more interesting, valuable and on-brand.