Researchers say health systems should be careful not to assume that an algorithm is unbiased, just because it's not human.
  • RubberDuck@lemmy.worldEnglish
    2·
    1 month ago

    Does that not depend on where you are doing the analysis? Because plenty of diseases are more and others less common in specific races. I also don’t know if diseases present different across races.

    But there is definitely a good warning that ai can be plenty biased if the dataset is biased… which historically it will be.

    • MoobythegoldensockEnglish
      4·
      1 month ago

      That’s what the article is about. Because race itself is not real (it’s a social construct humans made up,) race is working as a proxy in these situations. The push is to find the actual, objective factors that this subjective category is a proxy for, and then make decisions based on those objective things rather than using a proxy.