Singapore’s Model AI Governance Framework for Generative AI

  • 2024年5月31日 2024年5月31日
  • 亚太地区

Yesterday (30 May 2024), the Singapore government released its Model AI Governance Framework for Generative AI (“GenAI Framework”). This framework builds on the earlier Model AI Governance Framework and reflects emerging principles, concerns and technological developments in GenAI.

The GenAI Framework may be accessed here.

In this brief article, we have summarised the main points of the GenAI Framework below.

The GenAI Framework builds on the policy ideas highlighted in a discussion paper on generative AI jointly published by the Infocomm Media Development Authority of Singapore (“IMDA”), Aicadium and AI Verify Foundation and draws from insights and discussions with key jurisdictions, international organisations, research communities and leading AI organisations.

There are 9 dimensions which the GenAI Framework proposes to be looked at in totality, to foster a trusted ecosystem:

  1. Accountability – Putting in place the right incentive structure for different players in the AI system development life cycle to be responsible to end-users.
  2. Data – Ensuring data quality and addressing potentially contentious training data in a pragmatic way, as data is core to model development.
  3. Trusted Development and Deployment – Enhancing transparency around baseline safety and hygiene measures based on industry best practices in development, evaluation and disclosure.  
  4. Incident Reporting – Implementing an incident management system for timely notification, remediation and continuous improvements, as no AI system is foolproof.
  5. Testing and Assurance – Providing external validation and added trust through third-party testing, and developing common AI testing standards for consistency.
  6. Security – Addressing new threat vectors that arise through generative AI models.
  7. Content Provenance – Transparency about where content comes from as useful signals for end-users.
  8. Safety and Alignment R&D – Accelerating R&D through global cooperation among AI Safety Institutes to improve model alignment with human intention and values.
  9. AI for Public Good – Responsible AI includes harnessing AI to benefit the public by democratising access, improving public sector adoption, upskilling workers and developing AI systems sustainably.

The GenAI Framework sought to expand on the existing Model AI Governance Framework for Traditional AI (“Traditional AI Framework”). The 1st edition of the Traditional AI Framework was published on 23 January 2019 for broader consultation and adoption. The Traditional AI Framework provides detailed and readily-implementable guidance to private sector organisations to address key ethical and governance issues when deploying AI solutions. The 2nd edition of the Traditional AI Framework was released on 21 January 2020, where it includes additional considerations (such as robustness and reproducibility) and refines the original framework for greater relevance and usability.  The 2nd edition of the Traditional AI Framework may be accessed here.

Concluding Words

As generative AI continues to develop, there is a need for global collaboration on policy approaches. The GenAI Framework aims to provide a basis for global conversation to address generative AI concerns while maximising space for continued innovation. The ideas proposed seek also to further the core principles of accountability, transparency, fairness robustness and security.

The GenAI Framework is the first step towards fostering a trusted ecosystem for generative AI. Referencing the GenAI Framework’s 9 dimensions, the Singapore government will continue to engage key stakeholders to develop these guidelines and resources to enable a systematic and balanced approach towards building guardrails while enabling maximal space for generative AI innovation.  

Should you have any queries on the GenAI Framework and how this may affect you, please feel free reach out to the authors of this article and they would be happy to assist you.