Improving Kurdish (Kurmanji) grammar accuracy in ChatGPT using reliable sources
Hello OpenAI team and community,
As a native Kurdish speaker (Kurmanji dialect), I’d like to share an important observation and suggestion regarding Kurdish grammar generation in ChatGPT.
While the model’s ability to understand and generate Kurdish is appreciated, it still produces frequent grammatical inaccuracies—especially with word endings, noun formation, vowel-consonant transitions, and case suffixes.
To improve this, I highly recommend that OpenAI consider incorporating rules from a trusted linguistic reference:
“Rezimana Kurmancî” by Samî Tan — a well-respected grammar book used by Kurdish language educators and linguists across the region. It offers clear rules based on phonetics, morphology, and standard Kurmanji usage.
A specific example I encountered was the incorrect formation of “hêvîyek” instead of the grammatically correct “hêviyek”.
It would mean a lot to native speakers and educators to see accurate, respectful support for our language in AI tools, especially given the cultural and linguistic richness of Kurdish.
Thank you for your ongoing work and for supporting underrepresented languages like ours.