Speaking at the AI Women’s Forum during the 2026 World Artificial Intelligence Conference (WAIC 2026) held in July in Shanghai, Professor Yang Hongxia, Executive Director of the PolyU Academy for Artificial Intelligence (PAAI), Associate Dean (Global Engagement) of the Faculty of Computer and Mathematical Sciences, and Chair Professor of Generative Artificial Intelligence in the Department of Computing, delivered a keynote speech on technological equity that outlined a decentralised paradigm for an inclusive, AI-native economy.

A new paradigm for technological equity

Invited as a keynote speaker at the high-level AI Women’s Forum—co-hosted by the All-China Women’s Federation and the WAIC committee under the theme “AI Leap • Her Momentum”—Professor Yang delivered an address titled “Co-GenAI Platform: A New Paradigm of Technological Equity in the AI-Native Economy”. She highlighted that generative AI marks the beginning of a 10-to-30-year technological revolution, ushering in an “AI-native” economy where products, workflows, and organisations are designed around AI from inception.

Challenging the prevailing model, which is heavily reliant on massive capital and centralised compute, Professor Yang stressed that frontier AI must not remain “a game for tech giants and the wealthy”. Genuine technological equity requires overcoming key industry hurdles—such as data security constraints and prohibitive compute costs—to enable diverse organisations, countries, and individuals to truly master “their own data, their own models, and their own AI” within secure environments.

Empowering nations and diverse industries

To lower resource barriers, Professor Yang’s team has pioneered the Collaborative Generative AI (Co-GenAI) framework and the “Model-over-Models” (MoM) approach, underpinned by end-to-end FP8/FP4 low-bit training and the InfiFusion model fusion pipeline. This approach dramatically reduces model training costs to less than half those of conventional solutions.

Beyond cost reduction, this decentralised framework enables cross-border and cross-sector collaboration without requiring parties to share raw, sensitive data. By connecting specialised domain models trained on proprietary data, it creates a collaborative global intelligence network, which enables the following:

National Sovereignty: It allows developing nations to train dedicated “sovereign AI models” locally at affordable costs, protecting native data sovereignty while bridging the global AI divide.

Industrial Transformation: It empowers vertical sectors—such as manufacturing, healthcare, and finance—to build bespoke models. Speaking also at WAIC’s healthcare AI forum, Professor Yang illustrated how hospitals can train specialised models on local clinical data and fuse insights across departments like oncology and cardiology while preserving clinical data privacy.

Translating research into industry impact

The vision outlined at WAIC reflects years of research and industrial innovation led by Professor Yang and her team. Professor Yang brings a distinguished track record spanning academia and global tech giants. Holding a PhD in Statistics from Duke University, she formerly led foundation model development at ByteDance US and Alibaba Group, authored over 150 papers, and holds more than 50 patents. Her honours include the WAIC Super AI Leader (SAIL) Award and Forbes China Top 50 Women in Science and Technology.

To bridge laboratory breakthroughs with enterprise needs, Professor Yang founded InfiX.ai, a PolyU academic-led AI startup. The venture provides enterprise-grade AI infrastructure, including its InfiXspeed platform and InfiCube—an all-in-one training and inference appliance that won the prestigious BEYOND Best of Innovation Award 2026.

Unlocking female momentum in the AI era

Professor Yang underscored that generative AI acts as a great equaliser of knowledge, significantly lowering the technical barrier for women to engage in scientific research, engineering, and tech entrepreneurship. Drawing from her own interdisciplinary journey, she encouraged aspiring female researchers to maintain an open mindset and embrace continuous learning.

“Artificial intelligence does not just empower women—it needs more women to actively shape its future,” Professor Yang emphasised.