Brunei doesn’t just need another AI talk; it needs a workable plan. So at DevFest Brunei 2025, GDG Brunei, MKOKU, and the brunei_4AI Community pushed past the buzzwords and sat people at the same tables—developers, NGOs, policy folks, civil servants, and volunteers—to do the messy work. No speeches, no hype. Just a hard look at how our data is actually collected, shared, and protected today, and what we’re willing to change tomorrow.
DevFest Brunei 2025 hosted a joint tabletop exercise on AI and data governance, co-organised by Google Developer Group (GDG) Brunei, MKOKU, and the brunei_4AI Community. It wasn’t a slide-show about buzzwords. It was a working session to face uncomfortable truths about fragmented data, test practical ideas with stakeholders, and agree—together—what to do next.
The room started from first principles. Brunei’s data sits in silos. Ministries keep separate systems, methods aren’t standardised, and people “disappear” from databases when they leave one agency’s scope. That breaks follow-up, weakens service delivery, and introduces security risks. The deck put it plainly: no central repository, inconsistent collection, manual processing, and incomplete verification protocols (tracking and security gaps were called out as well).
Rather than debate abstractions, participants worked in five stakeholder tables—Government, NGO, Policy, Tech & SMEs, and Volunteers—and documented the realities they live with. The Government table flagged inter-agency distrust and pushed for a formal data governance framework, stronger access control and encryption, public–private collaboration, and capacity-building. NGOs highlighted the volunteer NDA gap and the painful “re-submit your documents” loop, proposing automated, consent-based updates. The Policy group called for a minister-level workshop and a common data model across agencies. The Tech & SME table pressed for top-cover from ministers, digitising ground inputs like Ketua Kampong records, and a lifelong “People ID” as an anchor for interoperability. Volunteers stressed integrity: quality assurance of datasets, regular pen-testing, a vetted national volunteer pool, and a sovereign cloud for sensitive merges.
Once ideas were on the table, the whole room voted using a simple three-lens test: which solutions would be most impactful if successful, which would be hardest to execute, and which were most urgent to implement. Everyone got three votes per poll, forcing tough trade-offs rather than feel-good consensus.
What surfaced from the voting and discussion was refreshingly pragmatic. The group did not chase shiny tools first; they prioritised foundations:
Leadership and rules before code. Minister-level sponsorship and a national governance framework were repeatedly named as prerequisites for progress, not nice-to-haves. Without shared rules, APIs won’t matter and pilots will stall.
A common model for people and data. The “People ID” concept kept resurfacing as the spine for interoperability and consented sharing, paired with a common data model that agencies can map into. It’s less about a plastic card and more about one durable identity that other datasets can reliably attach to.
Security and integrity as everyday practice. Quality gates on input, breach-minded design, and routine pen-testing were treated as table stakes, not afterthoughts. So was the case for a Brunei-hosted sovereign cloud for sensitive merges and auditability.
Ground truth, digitised. Ketua Kampong inputs are vital but still paper-based and scattered. Getting that data stream digitised—with guardrails—was seen as a high-leverage fix that makes any national system more representative and timely.
Human-AI workflows, not AI autopilot. Participants leaned toward AI agents that support a human-in-the-loop model for case management, verification, accessibility, and unstructured data—embedded into a secure, standards-driven platform rather than glued on later.
The exercise also put some hard truths on the table. Brunei’s current process relies heavily on manual spreadsheets and volunteer verification, which is generous but fragile. It creates inconsistent entry, version drift, and exposure to misuse. The cost isn’t just operational; it means vulnerable people are under-counted, and resources are allocated using partial pictures.
The takeaway is simple and demanding. If Brunei wants trustworthy, AI-ready public services, we must fix the plumbing: common standards, real leadership, a living “People ID,” security by default, and digitised ground truth. None of that is glamorous, all of it is doable. The partners behind this workshop will keep the momentum going with open notes, working groups, and small pilots that ship. If this matters to you—whether you write code, set policy, run a service, or just care about getting help when you need it—join the brunei_4AI Community. Bring your perspective, roll up your sleeves, and help build the next round of practical wins. We’re not waiting for perfect; we’re moving forward together.