Charlie StrossSep 24, 2025

Why LLMs can't handle everyday social etiquette in Persian:

Reading between the lines, this is a really good caution for any diplomats who are thinking of firing their interpreters and replacing them with "AI".

https://arstechnica.com/ai/2025/09/when-no-means-yes-why-ai-chatbots-cant-process-persian-social-etiquette/

When “no” means “yes”: Why AI chatbots can’t process Persian social etiquette

New study examines how a helpful AI response could become a cultural disaster in Iran.

Ars Technica
Show threadTom Bortels

@cstross

To be 100% fair - according to the article, even cultural native human speakers fail this sort of dance 20% of the time. I wonder how often a non-native human fails the same test?

There are lots of really good reasons to distrust a LLM - but "fails in situations where humans also fail frequently" isn't actually surprising, I hope.

Most tools have inappropriate usages. And human culture varies wildly.

Sep 24, 2025 at 1:02pmWeb
Show thread�Sep 24, 2025
@tbortels @cstross yeah, what i've thought, too. like how many western tourists fucked up in the taxi