[LINK] Data-Driven Understanding of Real-Life Moral Dilemmas

[Brendan]

On 18/4/23 10:50, David wrote:
> On 18/04/2023 09:34, Tom Worthington wrote:
>> [quoting the author...]  we operationalize the moral foundations theory—which proposes five fundamental dimensions of morality—in analyzing content on Reddit and Twitter
> That seems like a good example of garbage in, garbage out.

The whole of [currently fashionable LLM based-] "AI" is the distillation of biases latent in the training data into a formal structure. I have little experience with twitter but what I have seen of reddit suggests it has the moral sophistication of a 15 year old.