Continued Frontier AI Access for Singapore Given US Order to Bar Foreign Access to Anthropic's Fable and Mythos Models
Ministry of Digital Development and InformationSpeakers
Summary
This question concerns Singapore’s access to frontier AI models following US restrictions on specific Anthropic models, as raised by Mr Kenneth Tiong Boon Kiat regarding the government's strategy for securing access or alternatives. Minister for Digital Development and Information Mrs Josephine Teo replied that while frontier models are useful for niche research and cybersecurity, most current needs are met by already available models. She explained that the government will prioritize resilience by diversifying its range of AI models through partnerships with multiple providers to mitigate geopolitical and export control risks. Additionally, Singapore will advocate for an open AI ecosystem and collaborate internationally on AI safety and governance to preserve flexibility in meeting its technological requirements. This policy focuses on maximizing accessible capabilities while maintaining a multi-pronged approach to ensure the nation can meet its diverse AI needs.
Transcript
55 Mr Kenneth Tiong Boon Kiat asked the Minister for Digital Development and Information following the US order barring foreign access to Anthropic's Fable and Mythos models (a) whether the Government assesses that Singapore needs continued frontier AI access; (b) if so, what is the Government's strategy to secure such access against further US restrictions; and (c) if not, what alternatives will maintain sufficient capability, and what is the acceptable lag behind the frontier.
Mrs Josephine Teo: Access to frontier models is helpful for specific use cases, such as advanced research and cybersecurity. However, these form a small proportion of artificial intelligence (AI) demand. For most industry, Government and research uses, capable models are already available.
For the foreseeable future, access to frontier AI will continue to be shaped not just by commercial considerations but also export controls and geopolitics. We should therefore be realistic and focus on making the most of the considerable capabilities that remain accessible. At the same time, for better resilience, we will diversify and maintain access to a range of AI models through partnerships with multiple providers. At international fora, we continue to advocate for an AI ecosystem that is open and collaborate with partners on different aspects of AI governance, including safety, security and the responsible use of AI. This will help preserve flexibility to choose amongst available options to meet Singapore's AI needs.