Singapore - National AI Programme

Overview

AI Singapore (AISG) is a national programme announced on 3 May 2017 by the National Research Foundation (NRF) to catalyse, synergise and boost Singapore’s artificial intelligence capabilities. Its design combines funding, multi-agency governance and industry–research partnerships to pursue applied AI projects, foundational research and talent development. The programme set out concrete targets such as delivering a large portfolio of applied projects (notably the "100 Experiments" initiative), building a local talent pipeline (e.g., the AI Apprenticeship Programme) and creating shared resources like AI tooling and datasets. Official programme launch materials and NRF documentation set out the initial investment envelope (up to S$150 million over five years) and the management committee structure. Key official descriptions of the programme can be found in the NRF/Prime Minister’s Office announcement and contemporary government briefings, which frame AISG as a national enabler in Singapore's broader smart nation and AI strategy ecosystem. For the original press release see AI.SG: New National Programme to Catalyse, Synergise and Boost Singapore's Artificial Intelligence Capabilities (MDDI) and the corresponding NRF release NRF press release (2017 PDF).

Definitions

Key terms used in the programme and related governance documents include: "AI/AI systems" (applications and algorithms that perform tasks commonly associated with human cognition, including machine learning models); "translation" (the process of taking research outputs into deployable, production-grade solutions); "100 Experiments (100E)" (a structured portfolio of applied projects designed to rapidly develop and deploy AI solutions with industry partners); "talent pathways" (programmes such as AI Apprenticeship Programme and AI for Industry to educate and place AI practitioners); and "makerspace" (shared tooling, datasets and compute resources offered by AISG to lower adoption barriers). These working definitions are operational for AISG grants, MOUs and programme milestones and align to terminology used across Singapore’s AI ecosystem including IMDA, A*STAR and NRF guidance.

Governance and Institutional Framework

The governance model for AISG is multi-institutional and programmatic. The National Research Foundation (NRF) provides funding oversight and strategic sponsorship. A Management Committee composed of senior representatives from NRF, the Smart Nation and Digital Government Office (SNDGO), IMDA, EDB, research institutes and industry was set up to steer priorities and approve major initiatives. Operationally, AISG coordinates with implementing partners from universities and public research institutes (PRIs), while SGInnovate and other intermediaries assist in translating prototypes into commercial or public-sector deployments. Programme grants and projects are governed by contractual terms, deliverable milestones and reporting obligations; these set expectations for data governance, IP treatment and evaluation metrics. Governance also includes coordination with sector regulators where projects intersect regulated activities (for example, healthcare projects coordinate with the Ministry of Health and Integrated Health Information Systems). AISG explicitly references broader national frameworks such as Singapore’s National AI Strategy and collaborates with IMDA on governance and testing toolkits (e.g., AI Verify initiatives). For the IMDA collaboration and later LLM-related programmes see IMDA press release (Dec 2023) and A*STAR material acknowledging AISG’s role in national AI efforts: A*STAR press release (SEA-LLM).

Key Focus Areas

AISG’s mandate organizes activity across three interlinked pillars. First, applications and industry translation — delivering problem-driven projects (100E) in priority sectors (healthcare, finance, urban systems) to prove value and accelerate adoption. Second, research and deep capabilities — enabling advanced research (explainable AI, small-data learning and federated approaches) through fellowships and investigator awards to maintain scientific competitiveness. Third, talent and ecosystem — establishing training pipelines (AI Apprenticeship Programme, AI4I / AI4E curricula, PhD supports) and shared resources (makerspace, datasets, tools) to expand the pool of AI practitioners and lower barriers for SMEs. AISG also invests in tools, open-source releases and engineering best-practices to foster reproducibility and translation. Over time, AISG’s work has intersected with national-level initiatives such as the National Multimodal LLM Programme and SEA-LION models to ensure capability continuity and regional relevance.

Implementation Framework

Implementation is delivered through a mixture of direct grants, co-funding with industry, challenge-based procurements and collaborative MOUs. Projects follow a defined lifecycle: problem definition and stakeholder alignment; selection of principal investigators and engineering teams; development, evaluation and iteration; and deployment/translation with monitoring. AISG enforces milestone-based funding with defined deliverables and reporting requirements, and encourages open practices such as releasing anonymised datasets and reference implementations where permissible. Talent schemes mix classroom instruction with project-based practicum, embedding apprentices into industry projects to scale learning. The programme leverages national compute resources (e.g., NSCC partnerships) and works closely with IMDA on governance/testing toolkits for model evaluation and safety assurance.

Monitoring and Evaluation

Monitoring uses quantitative and qualitative KPIs: number of delivered projects and proofs-of-concept, commercialised solutions, talent placements and training completions, publications and open-source artefacts, and measurable productivity gains in partner organisations. Evaluation combines periodic reporting to NRF (financial and progress reports), independent technical reviews, and stakeholder feedback. AISG sets project-level risk assessments and requires mitigation plans; higher-risk deployments (e.g., in clinical or safety-critical settings) require additional validation and regulator engagement. The programme also measures ecosystem outcomes such as start-up formation, cross-institutional collaborations, and adoption rates among SMEs.

Penalties, Liability, and Appeals

AISG is an executive programme rather than statute; its enforcement levers are administrative and contractual. Typical remedies for non-compliance include suspension or termination of funding, requirement to repay grant sums (clawback), removal from AISG programmes, and reputational actions (public notice). Liability for deployed solutions rests with the executing parties and their contractual obligations; AISG’s grant terms typically require compliance with applicable laws, data protection rules (PDPC), and industry-specific regulations. Appeals against adverse administrative decisions are handled through formal programme appeal channels and contractual dispute resolution clauses (which may specify negotiation, mediation or arbitration). Programme documents emphasise adherence to ethical and legal standards but do not create independent criminal sanctions.

Relationship to Other Instruments

AISG functions as a programmatic delivery arm aligned to Singapore’s broader AI strategies and regulatory instruments. It complements the National AI Strategy (2019) and NAIS 2.0 (2023) by operationalising research-to-deployment pathways and talent development. AISG coordinates with IMDA’s governance frameworks (Model AI Governance Framework, AI Verify initiatives) and PDPC guidance on personal data protection. Sectoral guidance (for example, healthcare AI guidance from MOH/IHiS) is adopted at project-level when relevant. AISG’s open resources and testing collaborations support interoperability with national testing and conformity assessment efforts driven by IMDA and other agencies.

International Alignment

From inception, AISG has signalled international engagement: collaborating with overseas research institutions, participating in regional model development (e.g., SEA-LION) and aligning practice with international best-practice on AI governance, testing and safety. AISG’s ecosystem approach and emphasis on explainability, shared datasets and engineering toolkits are consistent with cross-border standards and initiatives on trustworthy AI. The programme has worked alongside multilateral partners and leveraged international vendor and academic collaboration to ensure Singapore’s outputs are interoperable and compliant with evolving global norms.

Implementation Timeline

Sources and References