Australian researchers have joined Project Tapestry, a global effort to build frontier AI models that harness shared knowledge while strengthening data ownership and AI sovereignty.
Monash University has become the first Australian organisation to join a pioneering international initiative established to bring together leading research and technology capabilities to develop open, trusted and sovereign artificial intelligence (AI) using a shared knowledge approach.
The Monash AI Institute has joined Project Tapestry, founded by US-based non-profit organisation The AI Alliance, guided by its Chief Science Advisor and Turing Award winner Professor Yann LeCun, who is Silver Professor at NYU’s Courant Institute.
The Project aims to develop a large-scale, advanced foundation AI model at the frontier of AI development by integrating trusted domain-specific models developed by its global members.
The partners will retain ownership of their data and compute, while benefiting from the shared model to develop their own sovereign AI systems within a distributed and collaboratively governed framework.
The Monash AI Institute is also closely collaborating with BharatGen, India’s sovereign AI initiative and a Project Tapestry partner, to both support the development of its AI models and also draw on expertise from India’s contribution to the global consortium.
Monash AI Institute Director, Professor Shonali Krishnaswamy, said joining Project Tapestry would strengthen Australia's ability to develop AI that reflects its own priorities and diverse communities while connecting Australian researchers with trusted expertise and resources internationally.
“Building sovereign AI does not mean building AI in isolation,” Professor Krishnaswamy said.
“Australia needs the capability to develop and shape AI here, while working with trusted international partners to build models that reflect our values, cultures and priorities.
“Through Project Tapestry, we can contribute Australian research and expertise to a global effort while retaining control of the data, knowledge and infrastructure that underpin our own AI capability.”
As one of its first contributions to the Project, the Monash AI Institute is planning to develop a socially and culturally aware trusted AI model to combat gender-based violence.
The model will draw on multidisciplinary expertise spanning AI, sociology and criminology led by researchers from the ARC Centre of Excellence for the Elimination of Violence against Women (CEVAW). The team aims to release the model with an accessible interface by the end of 2026.
CEVAW Director, Distinguished Professor Jacqui True said the model would address a critical challenge by ensuring information about gender-based violence is grounded in trusted evidence, rather than biases and harmful stereotypes that can exist in broader datasets.
“AI systems can reproduce gender biases, misogyny and spread misinformation contained in the data they learn from,” Professor True said.
“This becomes particularly concerning when people are seeking information about something as complex and sensitive as gender-based violence.
“We want to build a model grounded in trusted evidence that recognises how experiences of violence can differ across cultures and communities, and provides people seeking help as well as researchers and policymakers with information they can rely on.”
Once developed, this model could also be applied in diverse areas such as supporting financial inclusion and assessing lending to sections of society who may be disadvantaged by AI biases including ethnic, religious, linguistic minorities and women entrepreneurs.
As the first Australian partner to join Project Tapestry, the Monash AI Institute is taking a position of global leadership in strengthening Australia’s and the Indo-Pacific's sovereign AI capability. Central to this position of leadership is Monash’s Australia-first MAVERIC supercomputer, which will be used to build the AI model to combat gender-based violence.
Machine learning and natural language processing expert Professor Reza Haffari, who is leading Monash’s development of foundation models, said the project would demonstrate how trusted, domain-specific AI models could be built using secure and culturally relevant sources of knowledge.
“The challenge is not simply to build more powerful AI, but to build AI that people can trust,” Professor Haffari said.
“By combining Monash's research expertise and MAVERIC's computing capability with trusted data, our own extensive research in Natural Language Processing and international collaboration through Project Tapestry, we can build models that are both technically capable and grounded in the communities they are designed to serve.”
AI Alliance Chairman, Dr Anthony J. Annunziata said Monash would bring important Australian research capability and expertise to the growing international initiative.
"AI should belong to those who depend on it, and reflect their goals, knowledge and culture," Dr Annunziata said.
“Project Tapestry is about creating an open and collaborative approach where organisations can contribute to advanced AI development while retaining sovereignty over their data and models. Monash brings formidable AI research capability and an important Australian perspective to this effort.”
Project Tapestry recently reached its first technical milestone with Milestone M0 providing the first proof point for its Australia-India leg.
“As a part of Project Tapestry’s first technical milestone, Monash and BharatGen trained a shared model using sovereign computing infrastructure in Australia and India, while keeping all training data securely within each country,” Dr Annunziata said.
“It shows that sovereign AI can be built collaboratively, with Monash playing a key role in demonstrating how this can work."
Monash AI Institute Director Professor Shonali Krishnaswamy, AI Alliance Chairman Anthony J. Annunziata and BharatGen Principal Investigator Professor Ganesh Ramakrishnan are available for interviews.
Learn more about Project Tapestry:
Website: https://thealliance.ai/projects/tapestry
GitHub: https://github.com/The-AI-Alliance/tapestry
Data: https://thealliance.ai/projects/tapestry/training-data-proposals
About the AI Alliance
The AI Alliance is a global nonprofit research and technology organization dedicated to advancing open, safe, and responsible AI through innovation, collaboration, and advocacy. Operating through both a 501(c)(3) public-benefit organization and a 501(c)(6) innovation association, the Alliance brings together more than 200 collaborating organisations across 29 countries spanning industry, academia, startups, research, and government. The Alliance supports open projects and initiatives across AI data, models, agents, safety, and governance.
- ENDS -
MEDIA ENQUIRIES
Teju Hari Krishna
Media and Communications Manager
T: +61 450 501 248
GENERAL MEDIA ENQUIRIES
Monash Media
T: +61 3 9903 4840
For more experts, news, opinion and analysis, visit Monash News.
