By Seth Hays
The release of China’s DeepSeek R1 generative AI model, purportedly built on significantly less computing power and money than comparable models, made headlines across the world. While much discussion is focused on how this moves US–China AI competition, the rest of Asia is assessing the impact of this new player on their AI governance and industrial policies from practical and rights-based perspectives.
Examining policy moves by Japan, South Korea and Taiwan in response to DeepSeek shows the importance of AI governance. Particularly in the areas of privacy protection and national security, AI governance is necessary for a successful cross-border AI industry to thrive in Asia and positioning these countries to take advantage of industry volatility.
South Korea’s Personal Information Privacy Commission suspended the use of DeepSeek in South Korea due to concerns about the company’s collection and processing of personal data. The privacy regulator indicated that DeepSeek expressed its interest in actively cooperating with the authorities by appointing a local agent.
South Korea’s approach has been balanced in that it noted that the deficiencies can be remedied and the services restored after improvements are made. It cited similar concerns that were raised by US-based AI services, such as OpenAI and Google AI services.
South Korea is not favouring US-based AI over China’s tech sector, despite being a US ally, and privacy rules for AI companies are applied in a geopolitically non-discriminatory way. But the government’s actions do underscore the important role that AI governance will play in the country’s AI sector.
This follows the passage of South Korea’s AI Basic Act, which contains provisions on AI governance, such as requirements for transparency, accountability and a risk-based tiered system of regulation. Such legislation is aligned with the growing international consensus on trustworthy AI, such as the OECD framework. At the same time, it promotes industrial policy to foster the AI industry at home and support the creation of more data centres, accessible training data and education for AI talent.
In Taiwan, the Ministry of Digital Affairs imposed a moratorium on the use of DeepSeek in the public sector, citing national security concerns about data leakage. The concerns are bolstered by reports that DeepSeek spouts Chinese Communist Party propaganda regarding the status of Taiwan as an independent democratic society.
Though Taiwan has blocked other Chinese apps from being used by the public sector, its overall policy regarding AI and algorithmic technology from China has been based on principles supporting freedom of speech and access to technology. Taiwan has not blocked TikTok, despite concerns from civil society that the app could be used to manipulate public opinion or spread disinformation. Taiwan also expects to pass an AI Basic Act, which will institutionalise the need for AI governance processes to adhere to international standards on trust, accountability and transparency.
In Japan, Prime Minister Shigeru Ishiba has taken note of the implications of DeepSeek’s release as a reason to continue moving towards the passage of an AI Basic Act. He said that legislation on AI, expected to come in 2025, should boost the technology while minimising risks and dangers. Tokyo banned the use of DeepSeek on government devices, as did numerous companies, citing data security and privacy concerns.
Should the claims about DeepSeek’s compute intensity be accurate — and doubts exist about the very low price tag — this also spells an opportunity for many countries in Asia to enter the AI race without access to the advanced chips available in larger economies.
But the moral of the DeepSeek story is that ignoring AI governance, including data security and privacy concerns, is not a recipe for long-term success. Countries that foster trustworthy AI, respect privacy and data protection and do not have objectionable political bias will be able to compete long-term across borders.
Policies that ensure data and AI governance rules are interoperable will be key for building internationally competitive AI.
Japan has led the way in developing rules for data free flow with trust at multilateral venues, such as APEC and the OECD. South Korea has taken the lead in global AI governance discussions. In 2024, it hosted the second iteration of the AI Safety Summit and the Responsible AI in the Military Domain Conference.
Taiwan’s importance to global AI efforts cannot be understated, given that it produces the vast majority of the world’s advanced chips. The current administration has emphasised the importance of Taiwan to the ‘Democratic Supply Chain’. Taiwan is also considering an AI Basic Act to foster responsible AI development and use.
With the right policies, Japan, South Korea and Taiwan are well positioned to boost their domestic AI industries, despite regional volatility and geopolitical tension between the United States and China. As low-cost providers like DeepSeek continue to pick up steam, AI governance — complemented by rational infrastructure inputs — is crucial for a democratic AI supply chain.
About the author: Seth Hays is Managing Director and Co-founder of APAC GATES, a non-profit and association management consultancy based in Taipei.
Source: This article was published by East Asia Forum