121
'Don't use AI detectors for anything important,' says the author of the definitive 'AI Weirdness' blog. Her own book failed the test (fortune.com)
submitted 3 years ago by to c/technology@lemmy.world
9 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 12 points 3 years ago (1 child)

Using AI with humans-in-the-loop is a fantastically powerful combination. We can solve many problems with smart system design that humans alone cannot achieve.

Source: PhD on automated animal censusing: we can build visual databases of individual animals for passive and long-term conservation work.

  • permalink
  • fedilink
  • source
  • hideshow 2 child comments
    •
  • [–] 24 points 3 years ago (1 child)

    Absolutely, but that’s not what she’s saying. She’s saying that the products that tout the capability to detect the usage of LLMs to cheat on essays and the like are really rubbish and give a lot of false positives.

    She mentions that they’re particularly inaccurate when it comes to English-as-second-language speakers, meaning foreign/exchange students are more likely to get marked as cheaters even though they might not be.

    I think the issue is that our education system is dated. Grades in general aren’t effective measures of knowledge, and they suck as motivators.

    AI tools aren’t going anywhere so the education system will need to find a way of working with them. It’s time to modernise.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •