Data Use for AI: Notes from Singapore's First Data Festival

🇸🇬 Last week I joined a closed-door roundtable on 'Data Use for AI' at the inaugural Singapore Data Festival - co-organised by IMDA, PDPC, and the OECD. The room brought together chief data officers, AI leaders, healthcare practitioners, regulators, and technologists. Co-moderated by Clarisse Girot (OECD) and Denise Wong (PDPC), it was one of the most candid conversations I've been part of on data governance in the age of AI.

Here are the insights that stayed with me. 👇

The friction is not lack of data 🔥

Organisations have data. They have AI models.

But: Can I use this for AI? Who's accountable if something breaks? Is the quality provable?

The shift we're seeing is from 'data sharing' (one-off, static) to 'data access for AI' (continuous, governed, multi-party). That demands new infrastructure - not just new contracts.

Privacy-preserving computation has moved to production 📦

We're past the 'do PETs work?' question.

The conversation now is: what does it take to operationalise them at scale? Privacy-enhancing technologies are no longer a research curiosity - they're becoming the infrastructure layer that makes responsible data access possible.

Agentic AI changes the consent equation 🤖

When a user consents to data sharing, they're not necessarily consenting to an AI agent deciding how that data gets analysed, combined, or acted upon.

It doesn't map cleanly onto multi-agent autonomous flows. This is an open governance question - and one we need to answer together.

Data increases in value when shared responsibly 🌍

The challenge: building the infrastructure so data moves safely to where it creates the most value.

That means trust by design. Not as a compliance checkbox, but as the foundation of how we architect data flows in a world where AI agents are collaborators, not just tools.

Looking ahead 🚀

Grateful to be in the room. Thank you Clarisse Girot, Denise Wong, Johnson Poh, IMDA, PDPC, and the OECD team for convening this.

These conversations matter. They shape how we build AI systems that amplify human judgment - systems that are trusted, governed, and designed for the long term.

Looking forward to the next one. 🙏

#AlwaysDay1 #DataGovernance #AIGovernance #AgenticAI #TrustedAI #Singapore

Gallery