Policy Workshop

Fast-track your move to Frontier AI security governance

2 intensive days in Singapore for senior cybersecurity and digital governance officials: the people who brief ministers on cyber risk, represent their countries in multilateral fora, and are now making calls on agentic AI systems. Held during Singapore International Cyber Week (SICW) 2026, CSA's flagship cybersecurity event.

15–16 October 2026, Singapore. Closed cohort, by invitation only.

Format

2 days, in person

Location

ASEAN-Singapore Cybersecurity Centre of Excellence (ASCCE)

Dates

15–16 October 2026

Cohort

16–25 participants, by invitation

Organised by

Why this, why now

Frontier AI outpaces the rules

Agentic AI systems can now plan and sustain multi-stage cyber operations with minimal human direction. That changes the threat landscape faster than most governance frameworks have caught up with.

The officials who brief ministers on cyber risk, manage the traditional cyber threat landscape, or represent their countries in multilateral fora are increasingly the ones making critical decisions on agentic AI, often with minimal direct experience of how these systems actually work.

An agentic AI incident affecting a port, a payments system, or critical infrastructure doesn't stop at national borders. The officials who'd have to manage that kind of crisis don't yet share a playbook, consistent terminology, or a common threat model. This programme builds that shared ground in one room, over two days.

Who it's for

Built for senior officials, not data scientists

This track is for senior-level digital governance and cybersecurity officials who oversee AI security policy. It doesn't assume a background in data science or machine learning: the aim is to give you what you need to participate fully in technical discussions, not to turn you into a practitioner.

It suits people working in:

Cyber Policy & Regulation

You set or advise on AI and cybersecurity policy inside government, a regulator, or an international body.

Digital Governance & Critical Infrastructure

You oversee AI security policy for national digital infrastructure or a critical sector.

Risk & Operational Practice

You sit at the intersection of policy and operational cybersecurity practice.

International & Regional Coordination

You represent your country in multilateral or regional cyber fora.

This isn't an introduction to cybersecurity policy. It assumes you already carry that responsibility and want the technical grounding underneath the decisions you're making.

Who's in the room

16–25 officials, deliberately drawn from across the region

The cohort brings together senior digital governance and cybersecurity officials from Singapore and the wider region: Singapore government agencies alongside ASEAN and broader Indo-Pacific counterparts. The mix is deliberate: the point isn't only shared learning, it's the working relationships a real cross-border incident would actually require.

What you'll do

Two days, from threat to toolkit to a concrete plan

Day0

Pre-work (before you arrive)

A one-page AI basics primer is shared in advance, along with a short template: each participant arrives having named 2–3 known policy gaps or live AI initiatives from their own agency, plus one AI-cyber incident from their jurisdiction.

Day1

AI foundations, attack surface & frontier risk

You start from how modern AI actually works (data, training, inference, deployment) and where a static model becomes an autonomous agent. From there: the AI supply chain and which assets actually matter (training data, model weights, compute, APIs, and who holds each one), then the attack surface (poisoning, adversarial attacks, exfiltration, prompt injection) mapped against RAND security levels. You work through AI as a tool for attackers, for defenders, and as a target itself, and where today's defences run out. The day closes on a frontier-risk case study: a walkthrough of the HuggingFace incident, where you name the misalignment, loss of control, and misuse in what happened, then work through the open-weight vs. closed-source trade-offs governments actually have to decide on.

Day2

Policy levers, local context & memo

A policy keynote and working session on the levers actually available to you (standards, procurement rules, vulnerability disclosure regimes, capacity building, international agreements), grounded in a worked example. You then run a jurisdiction diagnostic: structured group work mapping your own country's offensive exposure, defensive rollout, existing coverage, and gaps. The day ends with you drafting a one-page memo (risk, owner, existing coverage, enforceable ask, urgency) plus an action list naming who to brief and what has to happen first, and stress-testing it in structured small-group feedback rounds before a closing plenary.

What you'll leave with

Judgement, a plan, a map, and a network

A working grasp of how modern AI and agentic systems function, their supply chain, and their attack surface.
A clear view of the policy levers available and which fit your national context.
A one-page memo with a concrete, enforceable ask, stress-tested against peer feedback.
A working relationship with the counterparts across the region you'd need to coordinate with in a real incident.

A closed cohort

The Policy Workshop runs by invitation only: there is no open application for this round.