share on
This proposed framework takes into focus Generative AI's "potential for significant transformation", and is based on the principle that AI should be easy to digest, transparent, and fair.
Singapore's AI Verify Foundation (AIVF) and Infocomm Media Development Authority (IMDA) have developed a draft Model AI Governance Framework for Generative AI.
As shared in an announcement this week, the proposed framework comprises nine dimensions, and is based on the principle that AI should be easy to digest, transparent, and fair. Further, it offers practical suggestions that can be applied by model developers and policymakers as initial steps.
Here is a breakdown of the Model AI Governance Framework for Generative AI:
Dimension | Function |
Accountability | Putting in place the right incentive structure for different players in the AI system development life cycle to be responsible to end-users. |
Data | Ensuring data quality and addressing potentially contentious training data in a pragmatic way, as data is core to model development. |
Trusted development & deployment | Enhancing transparency around baseline safety & hygiene measures based on industry best practices in development, evaluation & disclosure. |
Incident reporting | Implementing on incident management system for timely notification, remediation and continuous improvements, as no AI system is foolproof. |
Testing & assurance | Providing external validation & added trust through third-party testing, and developing common AI testing standards for consistency. |
Security | Addressing new threat vectors that arise through generative AI models. |
Content provenance | Transparency about where content comes from as useful signal for end-users. |
Safety & alignment R&D | Accelerating R&D through global cooperation among AI Safety Institutes to improve model alignment with human intention & values. |
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. |
An in-depth explanation of each dimension can be found here.
This proposed framework is an expansion of the existing Model Governance Framework that was last updated in 2020, which only covered Traditional AI.
With Generative AI, the updated framework aims to make use of the existing consistent principles that are needed to create a trusted environment, and Generative AI’s potential for significant transformation, in which end-users will be able to use it confidently and safely, whilst simultaneously allowing space for cutting-edge innovation, AIVF and IMDA said.
It also aims to facilitate international conversations among policymakers, industry, and the research community, to enable trusted development globally.
It was added: "AI governance remains a nascent space. Building international consensus is key, as demonstrated by the successful mapping and interoperability of national AI governance frameworks between Singapore and the US, through the IMDA and the US National Institute of Science and Technology (NIST) crosswalk.
"The proposed Model Governance Framework for Generative AI takes this one step further by covering the latest developments in Generative AI. In turn, this will inform Singapore’s next steps, as we adopt a practical approach to maximise both trust and innovation."
READ MORE: What you should upskill in 2024 to stay relevant in the current workplace: Skills in Singapore
Lead image / 123RF
share on
