The UNU Macau AI Conference recently concluded, highlighting the growing global race for AI dominance. Professor Tshilidzi Marwala shared an anecdote about a large language model’s incorrect response regarding his grandmother, underscoring the challenges of accuracy in AI predictions. The conference, attended by over 500 participants from 30 countries, focused on shaping AI policy frameworks under the theme “AI for Humanity: Building an Equitable Digital Future.”
Despite overlapping recommendations for multi-stakeholder dialogue, the summit took place amid massive investments in AI infrastructure, expected to reach up to US$3 trillion by 2029. This surge is driving demand for semiconductors and energy sources, while governance efforts struggle to keep pace with innovation, a phenomenon known as the Silicon Valley Effect.
Keynote speaker Simon See addressed the role of falling technological costs in making AI more accessible, while also pointing out infrastructure and funding challenges. Panellists discussed the dual role of AI in climate change and its energy demands, noting the need for a balanced governance framework.
The conference also launched a UNU–Springer Book series to support research on AI-driven solutions. Professor Marwala emphasized the importance of embedding human values in AI systems to foster trust and mitigate risks. As AI reshapes labor markets, understanding these shifts is crucial for developing policies that maximize benefits and minimize harm. The conference aimed to serve as a platform for innovative ideas to address these evolving challenges.
