• kamenLady.@lemmy.world
      link
      fedilink
      English
      arrow-up
      18
      ·
      6 months ago

      Misaligned AI systems can malfunction and cause harm. AI systems may find loopholes that allow them to accomplish their proxy goals efficiently but in unintended, sometimes harmful, ways (reward hacking).

      They may also develop unwanted instrumental strategies, such as seeking power or survival because such strategies help them achieve their final given goals. Furthermore, they may develop undesirable emergent goals that may be hard to detect before the system is deployed and encounters new situations and data distributions.

      Today, these problems affect existing commercial systems such as language models, robots, autonomous vehicles, and social media recommendation engines.

      The last paragraph drives home the urgency of maybe devoting more than just 20% of their capacity for solving this.

      • schmorp@slrpnk.net
        link
        fedilink
        English
        arrow-up
        10
        ·
        6 months ago

        They already had all these problems with humans. Look, I didn’t need a robot to do my art, writing and research. Especially not when the only jobs available now are in making stupid robot artists, writers and researchers behave less stupidly.

    • David Gerard@awful.systemsM
      link
      fedilink
      English
      arrow-up
      14
      ·
      6 months ago

      you can tell at a glance which subculture wrote this, and filled the references with preprints and conference proceedings