Post #2597109
2026-05-10 12:20 UTC
@ngaylinn@tech.lgbt it’s a great example, isn’t it… will use it for this week’s lecture, with the supplementary discussion question: “so what, if anything, makes you confident that it is not expressing the same kinds of bias in your requested summaries and explanations of research articles?”.
Replies (1)
-
@yacc143@mastodon.social 2026-05-10 14:29
@UlrikeHahn@fediscience.org @ngaylinn@tech.lgbt Checking the output, questioning the output, at least the same way I do with human juniors. And yes, I feel way more comfortable when ChatGPT looks up the 27 EU MS on the EU Commission website than when it “remembers them from training data.” And naturally, the way LLMs work is completely alien to us, so the failure modes are different, especially if one anthropomorphizes them.