Southeast Asia

Singapore

Singapore’s public technical-safety record is strongest in open testing infrastructure and university-based methods: AI Verify Foundation / Moonshot, SMU agent-safety and model-security work, and NUS TrustLLM plus formal-methods framing.

Government posture

The most visible posture is assurance-oriented and tooling-centric rather than centered on a frontier-model developer.

Standards and evaluation

Project Moonshot gives Singapore a concrete public LLM benchmarking and red-teaming artifact; TrustLLM adds a peer-reviewed benchmark contribution.

Research ecosystem

Public technical work spans agent runtime enforcement, LLM backdoor benchmarking, trustworthiness benchmarking, and formal methods with LLM agents.

International engagement

The open-source and standards-facing design of Moonshot gives Singapore an outward-facing role in cross-border assurance and evaluation discussion.

Seed entities

Singapore AI Safety Institute

Singapore

Ai Safety InstituteTier 1

Publicly structured as a designation-plus-partnership model rather than a standalone statutory institute: DTC is the designated AISI and IMDA is the overall government agency.

Frontier Ai SafetyEvaluationTesting InfrastructureAi Assurance

Confidence: HighVerified: 2026-06-05

AI Verify Foundation

Singapore

Evaluation InstitutionTier 1

Singapore’s strongest public technical-testing ecosystem node, relevant here because it operates Project Moonshot for LLM benchmarking and red teaming.

Ai AssuranceEvaluationTesting InfrastructureStandards

Confidence: HighVerified: 2026-06-06

Digital Trust Centre Singapore

Digital Trust Centre

Singapore

Research GroupTier 1

NTU-based centre designated as Singapore AISI and serving as the technical backbone of the Singapore model.

Technical Safety ResearchAi AssuranceTesting InfrastructureCapacity Building

Confidence: HighVerified: 2026-06-05

Infocomm Media Development Authority

Singapore

GovernmentTier 1

Update IMDA with Singapore’s assurance and testing implementation roles, including A.I. Verify launch, the Global AI Assurance Sandbox with AI Verify Foundation, the Singapore AI Safety Red Teaming Challenge, and 2026 testing-standardisation activity.

Ai GovernancePolicy RegulationAi AssuranceInternational Dialogue

Confidence: HighVerified: 2026-06-06

AI Verify Testing Framework

Singapore

Policy ProgramTier 1Needs primary

Core assurance framework in Singapore’s public stack; should be separated from Singapore AISI itself.

Ai AssuranceEvaluationStandards

Confidence: HighVerified: 2026-06-05

Project Moonshot

Singapore

Evaluation PlatformTier 1

Open-source LLM evaluation toolkit that combines benchmarking and red teaming for LLMs and LLM applications.

EvaluationTesting InfrastructureRed TeamingAi Assurance

Confidence: HighVerified: 2026-06-06

Global AI Assurance Sandbox

Singapore

Policy ProgramTier 1

Singapore-linked assurance programme for deployed or deploying GenAI applications, represented as application-level assurance and market-building infrastructure rather than foundation-model safety testing.

Ai AssuranceCapacity BuildingEvaluationTesting Infrastructure

Confidence: HighVerified: 2026-06-06

AI Tester Accreditation Programme

Singapore

Policy ProgramTier 1Needs primary

Distinct third-party testing accreditation programme in Singapore’s assurance ecosystem.

Ai AssuranceCapacity BuildingEvaluationTesting Infrastructure

Confidence: HighVerified: 2026-06-05

Singapore AI Safety Red Teaming Challenge

Singapore

Evaluation PlatformTier 1Needs primary

IMDA-organized multicultural and multilingual GenAI red-teaming challenge focused on harmful bias stereotypes across Asian contexts.

Red TeamingEvaluationAi AssuranceTesting Infrastructure

Confidence: HighVerified: 2026-06-06

Jun Sun Group at Singapore Management University

Jun Sun research group

Singapore

Research GroupTier 1

Publicly relevant through agent-safety, backdoor, and LLM-safety work such as AgentSpec and BackdoorLLM.

Agent SafetyModel SecurityRed TeamingTechnical Alignment

Confidence: HighVerified: 2026-06-06

Jin-Song Dong Group at National University of Singapore

Jin-Song Dong research group

Singapore

Research GroupTier 1

NUS group relevant through formal methods with LLM agents, trusted AI, and the TrustLLM benchmark.

Technical AlignmentEvaluationAgent Safety

Confidence: HighVerified: 2026-06-06

Open questions

Need clearer public links between Moonshot modules and external testers or case-study pipelines.
Need more official evidence for AI Singapore or A*STAR multimodal safety work before inclusion.