3 ms·

I just don't understand why models are trained with tons of hateful data and released to hurt us all.

• mcherm 1y ago
  I am confident that the creators of these models would prefer to train them on an equivalent amount of text carefully currated to contain no hateful information. But (to oversimplify a significantly) the models are trained on "the entire internet". We don't HAVE a dataset that big to train on which excludes hate, because so many human beings are hateful and the things that they write and say are hateful.
  • amluto 1y ago
    We do have models that could be set up to do a credible job of preprocessing a training set to reduce hate.
  • aredox 1y ago
    Then... Stop. Stop doing this shit. Stop poisoning the well. Stop cultivating and amplifying and spreading everywherethis shit. Stop it.
• scarface_74 1y ago
  The WSJ trained it on “hateful data”
• accrual 1y ago
  > why models are trained with tons of hateful data Because it's time consuming and treacherous to try and remove it. Remove too much and the model becomes truncated and less useful. > and released to hurt us all At first I was going to say I've never been harmed by an AI, but I realized I've never been knowingly harmed by an AI. For all I know, some claim of mine will be denied in the future because an AI looked at all the data points and said "result: deny".