Anthropic signs an AI safety pact with Singapore, one of the world’s heaviest Claude users

Anthropic has signed a memorandum of understanding with Singapore’s government to work together on AI safety, security and model evaluation, and opened an office in the city-state, its fifth in Asia-Pacific. The agreement was signed on Wednesday on the sidelines of Singapore International Cyber Week.

Singapore is one of Claude’s heaviest users. In Anthropic’s latest Economic Index it ranks second of 121 countries for Claude.ai use relative to its population, at nearly six times what its size alone would predict.

What the agreement covers

The MOU was signed by Kai Fong Chng, Permanent Secretary of the Ministry of Digital Development and Information (MDDI), and Anthropic co-founder Ben Mann. Josephine Teo, Singapore’s Minister for Digital Development and Information, and Michael Sellitto, Anthropic’s head of APAC policy, witnessed the signing. It sets out three areas of cooperation:

• Safety and evaluation: technical exchanges with Singapore’s AI safety, security and assurance institutions on model evaluation, testing standards and incident response.
• Research: work on AI adoption and its impact, with Anthropic sharing research and data with the government, including from its Economic Index.
• Adoption: help for small and medium-sized businesses, AI skills training and support for the government’s goal of building AI capability among public officers. Anthropic also says it will offer technical and policy expertise as Singapore writes its rules for AI.

The text of the MOU hasn’t been published, and Anthropic’s announcement doesn’t say whether Singapore’s testers will see models before they are released. That is a live question: last month the White House asked Anthropic and OpenAI to keep their newest models away from Britain’s AI Security Institute until US officials had tested them first.

“Trust is critical”

Both signatories framed the deal around safety. Mann said:

Singapore has been a leader in responsible AI, with a long track record of investment in model testing and evaluation. We founded Anthropic on the belief that capability and safety have to go hand in hand.

Ben Mann, co-founder, Anthropic

Chng said the government wants people to be able to use AI with confidence:

Singapore’s approach to AI has always been to enable innovation while keeping pace with its evolving risks. Trust is critical for businesses, workers, and the public to use AI with confidence.

Kai Fong Chng, Permanent Secretary, Ministry of Digital Development and Information

The relationship is not new. In 2023 Anthropic joined the Generative AI Evaluation Sandbox run by Singapore’s Infocomm Media Development Authority (IMDA) and the AI Verify Foundation, and later worked with them on a red-teaming project in English, Mandarin, Malay and Tamil. Anthropic says it does similar technical work with safety institutes around the world, a theme Britain’s AI minister pressed this week when he told labs not to release what they can’t control.

Who in Singapore already uses Claude

Anthropic lists a long roster of local customers. The figures are its own and haven’t been independently checked:

• Government: more than 1,500 engineers and technical staff at the Government Technology Agency (GovTech) use Claude Code to modernise government systems, and the Cyber Security Agency of Singapore is exploring Claude for cyber defence.
• Finance: DBS, Southeast Asia’s largest bank, uses Claude for software development and day-to-day work, and GIC, the sovereign wealth fund, is rolling it out across its investment and corporate services teams.
• Tech: Grab’s BriX sales assistant runs on Claude and saves its sales teams about 40,000 hours a quarter; Sea uses Claude to speed up Shopee’s customer service chatbot, and thousands of its staff code with it.

Banks and other regulated firms can now run Claude with in-region inference on Amazon Bedrock, keeping data processing inside the region. Our guide to where Claude is available covers access country by country.

A team on the ground

Dale Finlay, Anthropic’s general manager for ASEAN, said customers in Singapore “hold AI to an extremely high standard, especially when it comes to reliability and security”, and that a local team would bring the company closer to them. The office is open and hiring.

Senior Anthropic leaders are in Singapore this week to meet customers, developers and policymakers. The company hosted a Builder Summit for more than 300 developers on Wednesday, after a meetup run by its local community ambassadors drew more than 700 builders earlier in the week.

Why it matters

Governments want a say in how frontier AI is tested, and deals like this give a small, AI-hungry country a direct line to one of the biggest labs. Singapore gets technical exchanges and data; Anthropic deepens ties in one of its busiest markets per head. How much Singapore’s testers actually get to see depends on terms nobody has published.

Sources: Anthropic: Anthropic opens Singapore office and signs MOU with the Government of Singapore.