SilkRoadNLP: A Workshop for the Iranian Language Family in NLP

Natural language processing has advanced rapidly in recent years. But much of that progress has focused on a relatively small set of well-resourced languages. Entire language families remain underrepresented—not because they lack linguistic richness, but because they lack visibility in data, tools, and research attention.

The languages of the Iranian linguistic family (Persian, Kurdish, Pashto, Balochi, Gilaki, Ossetic, and others) fall into this gap. They are widely spoken, culturally significant, and linguistically complex, yet there has been no dedicated venue for research on these languages within the NLP community.

SilkRoadNLP was created to begin addressing that gap.

As an ACL/EACL workshop, SilkRoadNLP met in March 2026 and brought together researchers working across these languages to share work, identify common challenges, and begin building a more connected research space. The workshop was a first step; the goal now is to grow an ongoing community and make these languages more visible within NLP.

You can learn more about the initiative and ongoing efforts here: [SilkRoadNLP Website] The workshop proceedings are available here: [Proceedings at ACL Anthology]

This is not just about expanding language coverage. It is about recognizing that linguistic structure, meaning and culture play a central role in how AI systems process and generate language. When entire language families are missing from the conversation, AI systems reflect that absence.

SilkRoadNLP is a first step toward building a space where these languages are not peripheral, but central to how we think about multilingual AI.