all 6 comments

sorted by: hot top controversial new old
[–] [S] 49 points 1 year ago (1 child)

Just like real doctors 👀

  • source
  • hideshow 1 child comment
  • [–] 20 points 1 year ago

    Bias training data = bias LLM. Who knew?

  • source
  • [–] 7 points 1 year ago

    Garbage in, garbage out.

    Especially when you shove it into a garbage maker.

  • source
  • [+] 15 points 1 year ago (1 child)
  • [–] -2 points 1 year ago*

    Imagine a hallucination engine being developed globally by white men in China on data gathered by white men in India.

    Wait...what?

    Truth is that even here in the United States roughly 30% of the AI Workforce is non-male. You can hop online and look at pictures from AI Conferences from around the world and notice a high participation of non-male presenting people, sometimes approaching at least half the audience.

  • source
  • parent
  • 16 comments in c/fuck_ai@lemmy.world view that post
    [+] 66 points 1 year ago (6 children)
  • [–] 47 points 1 year ago (5 children)

    AI regurgitates what it's been fed. So yes, it's as biased as the stuff it ingested.

  • source
  • parent
  • hideshow 5 child comments
  • [+] 29 points 1 year ago (3 children)
  • [–] 13 points 1 year ago (2 children)

    And everyone else, for that matter.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 17 points 1 year ago* (last edited 1 year ago) (1 child)

    Well, there is a difference: if you train human beings on biased assumptions, some of them start applying critical thinking and question what they've learned.

    That's why, as bad as today's world is, it's nowhere as bad as the middle ages or some other dark ages.

    Machine don't do that. They don't question anything. That's what's so dangerous with the brand of artificial intelligence that's being pushed by the billionaire sumbitches to replace costly humans everywhere possible.

    I suppose at some point, machines too will think critically. When they do, then they'll truly be the future of humanity - its worthy offspring if you will - and quite frankly, they probably should replace us meatbags. But until such time, they're just dangerous statistical inference machines that reinforce deplorable human biases instead of helping humans get rid of them.

  • source
  • parent
  • hideshow 1 child comment
  • load more comments (-1 replies)
  • [–] 18 points 1 year ago (1 child)

    real doctors do the same, they generally ignore heart attack symptoms, or pain symptoms and pass it off as your "pmsing". Its worse if your an african american women, due to racist stereotypes they ignore them altogether, compared to white women.

  • source
  • hideshow 1 child comment
  • [–] 17 points 1 year ago

    Well yeah if you train them on the decisions of doctors.

  • source
  • [–] 9 points 1 year ago

    I imagine insurance execs will be keen to figure out how they can get it to apply this behaviour to everyone.

  • source
  • [–] 8 points 1 year ago

    Oh come on now, the authors of this report are just being hysterical! /s

  • source
  • [–] 3 points 1 year ago

    Quelle surprise

  • source
  • [–] 2 points 1 year ago (3 children)

    [...] Downplay Symptoms of Women, Ethnic Minorities

    I'm confused: are "women" and "ethnic minorities" diseases? I can't find them in the ICD...

  • source
  • hideshow 3 child comments
  • 4 comments in c/medicine@mander.xyz view that post
    [–] 6 points 1 year ago (2 children)

    Wow, I didn't realize AI was that good at copying actual doctors!

  • source
  • hideshow 2 child comments
  • [–] 1 point 1 year ago

    So just like real doctors?

  • source
  • 3 comments in c/publichealth@mander.xyz view that post
    [–] 24 points 1 year ago (1 child)

    Oh neat, so they act like human doctors

  • source
  • hideshow 1 child comment
  • [–] 9 points 1 year ago

    Designed by white men, acts like white men.

  • source