Flint advises international businesses and investors on policy, politics, regulation and competition.
Flint advises international businesses and investors on policy, politics, regulation and competition.
Quick links
Sectors
Follow Us
AI-Pac is Flint’s digest providing our take on major AI policy and regulatory developments in Asia-Pacific. In a fast-moving environment, this newsletter is designed to help you stay on top of the regulatory agenda and consider what it means for your business. Please see here for our previous blogs, spanning 2024 to 2026. In each blog, we set out:
In July and August, AI policy across Asia-Pacific continued to evolve along several fronts. China strengthened its efforts to shape international AI governance. Meanwhile, Australia announced a national framework linking AI development and investment to strategic national priorities. Across the region, governments also strengthened their responses to AI-enabled cyber risks and clarified their approaches to the use of personal data and copyrighted works in AI development.
In this digest:
China is stepping up its ambitions to shape the institutions and rules governing the development and use of AI. The establishment of the World Artificial Intelligence Cooperation Organisation (WAICO) was the centrepiece of the 2026 World AI Conference in Shanghai (WAIC), with 29 countries signing its founding agreement. President Xi Jinping framed the organisation as a means of building a more “just and equitable” system for global AI governance, strengthening the Global South’s role in shaping AI. Xi also announced 5,000 AI training opportunities for developing countries and plans for AI application cooperation centres with groupings including ASEAN, the African Union and BRICS. This comes as increasingly capable and relatively inexpensive Chinese open-weight models, including Moonshot AI’s Kimi K3, gain international reach.
The combination of technology access, capacity building and institution building gives China a broader route to influence across both the technological and governance dimensions of AI. Beijing is seeking to embed Chinese technology and expertise within global South countries building their AI ecosystems while creating new forums through which those governments can participate in global governance. For global technology businesses, a key question is where and to what degree China’s efforts translate into sustained influence over technical standards, governance norms and technology choices.
This is sharpening the geopolitical choices facing APAC governments. The US-backed Pax Silica offers a contrasting model centred on trusted technology partnerships and supply chain security, with Australia, India, Japan, Singapore, South Korea, and the Philippines participating, while WAICO emphasises broad participation, capacity building and multilateral cooperation, including countries such as Indonesia, Malaysia, Cambodia, Laos and Myanmar. Reports that Washington is pressing countries to choose between them illustrate the increasing difficulty for APAC governments of maintaining flexibility in their positioning on AI. There is a growing possibility that geopolitical alignment will increasingly shape and restrict market opportunities for AI developers and deployers in individual countries across the region.
Australia took a significant step towards more centralised and mandatory AI governance in July, with Prime Minister Anthony Albanese setting out a new framework intended to ensure that AI investment and development advance Australia’s national interests. The government will establish a set of Australian Standards for AI, with supporting legislation expected in early 2027. A new Office of AI, established within the Department of the Prime Minister and Cabinet, will coordinate the design of the standards and bring together AI-related work across government.
The announcement reflects a more assertive view of the terms on which Australia should attract AI investment. Australia sees its land, renewable energy potential, skilled workforce and institutional stability as assets that can attract global investment, but increasingly expects that investment to contribute to wider national priorities. For large AI data centres, this includes expectations to underwrite new power supply, pay their full share of grid connection costs and fund additional water infrastructure where required. The Government’s decision not to introduce a broad text and data mining (TDM) exception for AI training reflects a similar instinct to balance the development of AI with the interests of Australian creators and rightsholders.
Australia’s approach could become an influential regional reference point for how governments seek to capture greater domestic value from AI investment. The “Canberra effect” is already visible in other areas of technology policy, with Australia’s world-first under-16 social media restrictions influencing debates and subsequent measures globally. Its AI framework could similarly shape how governments think about the conditions attached to AI investment.
Over July and August, governments across APAC strengthened their responses to the cybersecurity implications of increasingly capable AI systems. Japan's second AI Basic Plan calls for stronger defences against AI-enabled cyberattacks, including testing advanced AI models and inspecting vulnerabilities across critical government systems. Singapore issued an updated Cybersecurity Code of Practice for Critical Information Infrastructure. Indonesia’s draft Cybersecurity and Cyber Resilience Bill places greater emphasis on a “security by design” approach, embedding safeguards into technologies from the development stage.
Beyond responding to AI-enabled threats, governments are increasingly focused on building the AI capabilities needed for cyber defence. India is deploying AI for phishing detection and vulnerability assessment, while South Korea plans to launch a sovereign AI model specialising in cybersecurity by year-end. Japan and Korea are also placing greater emphasis on reducing reliance on foreign AI providers for sensitive applications. As AI becomes more important to cyber defence, questions around who provides the underlying models — and whether governments can depend on continued access to them — are becoming security considerations in their own right.
This could strengthen the push for sovereign or domestically controlled AI in some of the areas where governments perceive the greatest national security risk. Rather than seeking domestic alternatives across every AI application, governments may increasingly prioritise local capability in sensitive areas such as defence, cybersecurity, critical infrastructure, and government systems.
Governments across the region took further steps to clarify how privacy and copyright rules apply to AI training. Japan and Korea are creating greater flexibility around the use of personal data, easing some of the constraints facing developers seeking to train competitive models. Japan expanded the circumstances in which personal data can be used without consent. Korea similarly amended its privacy law to allow certain personal data to be used beyond its original purpose, subject to safeguards and regulatory approval, while separately launching a taskforce to reform the wider privacy framework for the AI era. Meanwhile, Australia moved in a more protective direction, proposing a new “fair and reasonable” test for the collection, use and disclosure of personal information, alongside stronger standards for consent.
Copyright policy presents a different set of choices, including whether to facilitate the use of protected works, strengthen licensing and transparency requirements, or rely on technical measures to protect rightsholders. Japan remains relatively development-friendly, adopting a voluntary “comply or explain” Transparency & IP Code to encourage disclosure of training data and methods. Singapore is reviewing whether its existing balance remains appropriate. Its public consultation considers whether to clarify its existing computational data analysis exception, the role of industry-developed technical measures and standards, and how responsibility for infringing AI-generated outputs should be allocated among developers, deployers and users. Australia, by contrast, has placed greater emphasis on protecting rightsholders, rejecting a broad TDM exception and stating that Australian creative works should not be used for AI training without the creator’s control.
Governments across the region are balancing competing interests and priorities as they consider their options on privacy and copyright. This includes weighing up the potential benefits of accelerated AI adoption against the potential implications for individual rights and for the collective interests of rightsholders. The result is likely to be greater clarity within individual markets, but continued divergence across the region. AI developers will face different rules on what personal data and copyrighted material can be used for training, what safeguards or permissions are required, and what degree of transparency is expected.
This blog was written by Ewan Lusty, Partner, Sarah Lim, Manager, and Nidhi Gupta, Senior Consultant. Over six years of working for Flint in Europe and Asia, Ewan has advised leading developers and deployers of AI on how to engage impactfully with governments on AI policy and regulation. In Sarah’s previous role at Singapore’s Ministry of Digital Development and Information, she coordinated whole-of-government policy positions on AI governance under Singapore’s National AI Strategy 2.0. Before joining Flint, Nidhi advised major technology companies and industry associations on India’s evolving digital policy landscape, including AI governance, competition policy, online safety and digital trade.