What?
The AmericasNLP 2026 Shared Task on Cultural Image Captioning for Indigenous Languages (CICIL) challenges participants to develop systems that generate accurate, culturally grounded captions for images depicting Indigenous cultures of the Americas, written in the Indigenous languages themselves.
Why?
Many Indigenous languages of the Americas are vulnerable or endangered, spoken by only small groups of people, and at high risk of extinction. Many communities are carrying out revitalization efforts, but creating teaching materials is expensive and time consuming. Effective language learning materials should also convey cultural knowledge, making culturally grounded visual resources particularly valuable. However, images depicting Indigenous cultures paired with descriptions in the Indigenous languages themselves remain scarce. NLP, and image captioning systems in particular, present an opportunity to help generate such materials at scale.
In addition to being endangered, most Indigenous languages of the Americas are so-called low-resource languages: the data needed to train any NLP systems, let alone deep learning-based systems, is severely limited. This means that many approaches used for high-resource languages, such as English and Chinese, are not directly applicable or perform poorly. Beyond data scarcity, generating accurate captions requires not only linguistic competence but also cultural knowledge, understanding the people, traditions, and contexts depicted in the images. The goal of AmericasNLP is to motivate researchers to take on the challenge of developing systems for these Indigenous languages.
Important Dates
Release of pilot data and baseline system: February 20, 2025
Release of development sets (50 examples): March 1, 2025
Release of surprise languages (if any): April 1, 2025
Release of test sets: April 20, 2025
Submission of results (shared task deadline): May 1, 2025
All deadlines are 11:59 pm UTC -12h (AoE).
Organizers (Alphabetically)
David Ifeoluwa Adelani, Minh Duc Bui, Luis Chiruzzo, Rolando Coto-Solano, Abteen Ebrahimi, Silvia Fernández, David Guzman, Elisabeth Mager, Manuel Mager, Rubén Manrique, Arturo Oncevay, John E. Ortega, Robert Pugh, Shruti Rijhwani, Raúl Vázquez, Katharina von der Wense
Contact
americas.nlp.workshop [at] gmail.com