When it comes to AI adoption, government is in no hurry. And that’s exactly the point.
In our latest AI survey, 50% of respondents working in the public sector said their organisation is still in the experimental or pilot stage of AI use. Compared to many private-sector industries, where early adoption is already shifting workflows and job design.
While at first glance, it might look like governments are falling behind, there’s good reason they move differently.
Why governments move slowly on AI
Government agencies aren’t built to “move fast and break things”. According to our in-house AI expert, Jack Jorgensen, General Manager of Data, AI & Innovation here at Avec, “There’s a big difference in the way governments need to operate versus private enterprise. They’re designed to be stable, reliable, and robust.”
A government body’s core responsibilities of public services, infrastructure, safety and regulation demand caution, reliability, and trust. So, when your ‘customer’ is the entire population, the stakes are high. Errors can impact millions, data breaches can threaten national security, and AI decisions must stand up to legal and public scrutiny.
The reality on the ground
In many agencies, AI is still in the exploratory stage:
- Small, controlled pilots
- Internal tools tested in low-risk areas
- Strong focus on compliance and security requirements
- Longer approval cycles for procurement and deployment
“Policy-making roles are challenging to automate and in highly regulated environments, finding relevant and safe use cases understandably takes time,” says Jack.
Security and compliance non-negotiables
Government respondents ranked “security and compliance concerns” on par with financial services and is no surprise given the sensitivity of the data they hold.
Some agencies are also grappling with:
- Lack of relevant applications – 20.2% said AI doesn’t apply to their current work
- Ownership uncertainty – it’s widely unclear who should lead AI initiatives
- Siloed operations – meaning slow cross-department collaboration
Why this pace makes sense
Jack says, “If anyone’s surprised government is slow on AI adoption, they don’t understand the role. The systems are meant to be dependable, not bleeding edge.”
While speed matters for the private sector in competitive markets, stability matters more than anything in public service. AI in government must work every time, be explainable and auditable, serve the public interest, and align with legislation and policy.
So, what can government do next?
- Continue piloting in low-risk and high-value areas
- Invest in AI literacy for leadership and frontline teams
- Create clear ownership and governance frameworks
- Learn from private-sector implementations without importing their risk appetite
- Build secure, compliant infrastructure before scaling
The private sector can afford to experiment, and government can’t, so caution at this stage isn’t failure. In an era where public trust is fragile, deliberate and well-governed AI adoption is the only responsible path.
Want to explore the sector-by-sector data? Access the full report.
If you want to plan a secure AI pilot, partner with Jack’s team at Avec.
