Leveraging collective expertise: AccelerateGOV Bitesize Insights on AI
As AccelerateGOV approaches on 9 December in Ottawa, senior leaders from Canada and abroad are focusing on one theme: making artificial intelligence useful, safe, and measurable in government. The goal is simple-share what works, skip what doesn't, and move faster together.
AI at pace, without cutting corners: Dany Bernier, RCMP
Dany Bernier, chief technology officer and cybersecurity lead at the Royal Canadian Mounted Police, is focused on scaling AI services across diverse ecosystems-responsibly, securely, and with clear value. He's pushing for a collaborative platform to share projects, lessons, and roadblocks across teams.
"By working transparently and leveraging collective expertise, we can increase the speed of delivery, reduce redundancy, and maximise the impact of AI across the organisation," he said.
- Catalog AI pilots across departments to spot duplication and reuse assets.
- Share pain points openly to help teams prioritise and move faster.
- Stand up an internal showcase for proofs of concept, models, prompts, and evaluation methods.
Strategy with real-world input: Jordan Zed, Government of Canada
Jordan Zed, senior assistant deputy minister at the Artificial Intelligence Secretariat (ISED), is helping shape Canada's new AI strategy with ministerial leadership. He's gathering perspectives from Canadian and international peers to stress-test what the federal government should prioritise next.
- Focus areas: meaningful adoption, public trust, measurable outcomes, and talent readiness.
- Bring back reusable frameworks, not just ideas-procurement models, guardrails, and patterns that scale.
Governance that travels well: Romina Kostani, Albania
Romina Kostani, director at Albania's National Agency of Information Society, is zeroing in on practical, ethical use of AI that supports national priorities. She's comparing how countries implement governance, risk classification, and accountability-highlighting Canada's work on responsible AI and the Algorithmic Impact Assessment, and risk models aligned with EU thinking.
She wants to ensure auditability and algorithmic transparency, so public institutions can explain decisions and uphold trust. Albania's own AI applications in services and policymaking will be part of the exchange.
- Use clear risk tiers and document decisions from day one.
- Adopt audit trails, model cards, and human-in-the-loop checks for high-stakes use cases.
- Learn from Canada's Algorithmic Impact Assessment and adapt it to local context.
From digitising to transforming: Carlos Santiso, OECD
Carlos Santiso, senior adviser at the OECD, is clear: digital tools alone aren't the goal. Government must be redesigned to deliver more personalised, proactive, and integrated services-especially under fiscal pressure.
He points to lessons captured in the OECD's work on governing with AI and Canada's leadership through the G7 GAIN. The priority now: prove benefits for communities with hard metrics, not headlines.
- Measure outcomes that matter-time saved for citizens, reduced backlog, faster benefits delivery, fewer errors.
- Simplify bureaucracy with smarter regulation and re-usable service components.
- Share what works and why, so others can replicate quickly.
Enable innovation safely: Dugald Topshee, Transport Canada
Dugald Topshee, chief service and digital officer at Transport Canada, is focused on shifting from "regulating innovation" to enabling it safely. With automated vehicles, remotely piloted aircraft, and complex marine operations moving fast, the old paper cycles won't keep up.
He wants to apply AI-powered digital twins to simulate scenarios, test compliance, and spot risks before they show up in the real world. "This would allow us to issue regulations, permits, and exemptions faster and with greater confidence."
- Use digital twins to test policy and compliance virtually before rollout.
- Define data, ethics, and governance standards upfront-especially for high-stakes systems.
- Share models and test results across agencies where mandates overlap.
Upskilling public servants for AI work
Topshee also emphasises talent. Inspectors and analysts will spend less time checking paper and more time validating models. That means AI literacy, data ethics, and prompt engineering now belong in core training for regulatory and program staff.
- Make Responsible AI training mandatory for non-technical roles involved in policy, service delivery, and oversight.
- Teach model limits, bias risks, red-teaming basics, and how to read evaluation reports.
- Give people safe sandboxes to practice with real tasks and datasets.
If your department is standing up learning pathways, you can shortcut the curation work with role-based course collections here: AI courses by job.
About AccelerateGOV
AccelerateGOV is co-hosted by the Government of Canada and brings together public servants to share practical insights on delivering better, more efficient, and resilient digital public services. The event features real-world use cases and lessons from nine countries and 58 speakers.
The first AccelerateGOV Bitesize Insights article - "Advancing transformation is a team sport" - explored interoperability, shared services, and digital sovereignty. The next Bitesize Insights article will be published on 8 December.
Where to find these speakers at the event
- Dany Bernier (RCMP): "Cyber security for all: Addressing the evolving threats".
- Jordan Zed (Artificial Intelligence Secretariat, ISED, Canada): "Data sovereignty in an age of AI and fragmentation".
- Romina Kostani (National Agency of Information Society, Albania) and Carlos Santiso (OECD): "How artificial intelligence can lead to greater productivity: Sharing best practice from the Government of Canada and beyond".
- Dugald Topshee (Transport Canada): "Boosting the health of government's technology".
What to take back to your team
- Stand up a shared AI registry to reduce duplication and speed reuse.
- Adopt simple, auditable risk tiers with human oversight for high-impact services.
- Measure benefits in citizen time saved, costs avoided, and service quality-report monthly.
- Invest in AI literacy for policy, regulatory, and service staff, not just data teams.
Your membership also unlocks:
