Capability 3.1.1.9
Culturally & Nationally Inclusive Models
Definition
Models trained on nationally significant datasets, including national language(s)/dialects, flora and fauna, and on weightings that encode local cultural, social and ethical values into the system, while maintaining Indigenous data sovereignty and guarding against exploitation.
Examples
- Matilda (Maincode)
- Early Indigenous language NLP projects (University of NSW, CSIRO)
- Domestic LLMs with AU English support such as Indigenous/heritage language models.
Indicators
- Number of domestically developed models
- Number of papers published on model development by researchers working for Australian entities
National AI Plan
PLAN COMMITMENT
Limited commitment
See 3.1.1.1 (Computer Vision (General)). The plan acknowledges importance of inclusive AI development and culturally appropriate datasets (Action 2) but no specific commitments. Limited commitment by government is a missed opportunity to increase domestic maturity in a globally scarce capability with high public interest value. It is particularly important to increase agency in this capability if investments are not made in General Purpose & Frontier Models (3.1.1.10).
Stocktake
Indigenous peoples and artificial intelligence: A systematic review and future directions. Perera et al. 18 June 2025, Sage Journals. https://journals.sagepub.com/doi/10.1177/20539517251349170?int.sj-full-text.similar-articles.6
Is the Stocktake missing evidence?
Share additional information for potential future Assessments