Trust in AI is on the fence, with KPMG data finding that only 46% of users are willing to trust what it creates. Given that there are approximately 20.2 million Gen AI users in the UK alone, that’s a lot of people still wary about the technology.
But for the public sector, which holds and manages a significant amount of sensitive data – from patient records and local education data to housing workflows and electoral data – trust cannot be taken for granted.
So, with AI use on the rise, how can those in the public sector increase trust in AI across their internal and external stakeholders?
Five trust factors
1: Capability – Boosting trust in AI starts with addressing literacy. The data suggest that the two are interlinked, and this is further reinforced in the public sector. The Productivity Institute found that 71% of UK public sector managers say workforce skills are the biggest AI obstacle. Turning this around requires organisation: considering what learning or upskilling is needed, planning how to do this for the benefit of teams, and adapting ways of working to embrace AI, all while aligning with the organisation’s culture.
As skills grow, teams will become what is considered an “intelligent customer”, whereby they hold enough in-house capabilities to conduct technical due diligence, evaluate vendor performance independently and negotiate contracts. This increased knowledge will also lead to organisations not being locked into any specific AI tools by design (which could, in the long term, harm trust in the technology).
2: Ethics by design – The ethical use of AI is high on the agenda for many, perhaps more so with the upcoming EU AI Act, which will be enforceable from August 2026. However, rather than view the ethical use of AI as a tick-box activity, organisations must embed it in their core model.
As AI processes and policies are built and established within an organisation, so too should the guardrails within which people operate. Additionally, just as an organisation would conduct project post-mortems, it should also run regular pre-mortems as AI is rolled out across departments to reason backward and review potential failure cases that could impact AI trust.
3: Humans in the loop – The next route to build trust is to reinforce the fact that any AI being embedded within an organisation or department is not there to replace, but to enhance.
At its core, any AI implementation must be human-centric, meaning that any active AI agent or automated workflow must have a named human owner who oversees its operations and controls its use. With clear intervention points and escalation triggers, human ownership should be embedded into the design of an AI process.
4: Cost – We’ve recently seen organisations pulling access to AI tools following high demand, resulting from huge token use. To drive trust and limit budget shocks, the public sector will need to match the architecture to the use case, treating FinOps as a discipline.
This is possible by establishing strict data foundations, which will ensure the success of any implementation. Again, this is where human oversight is vital to ensure budgets are managed appropriately.
5: Adaptability – While AI use cases mature and skills across the public sector grow, it will be important to build a level of adaptability into projects. This starts with small pilots or Proofs of Concept (PoCs) to test the validity and benefits of AI internally, in addition to using these as a basis for learning.
With continuous feedback loops, teams can create an implementation that, when it gets to the production phase, will be fit for purpose and run correctly the first time, reinforcing trust.
Establishing trust for a new tool or process never happens overnight, and when it comes to AI, which has long faced somewhat of a negative perception, this is more than true. As more workplaces, including those in the public sector, seek to embrace an AI future, building trust across teams is a must, and with ongoing feedback loops, transparency and flexibility, it will be possible for the sector to lead from the front in our AI future.
Are you ready to consider what your AI future looks like? Contact the CirrusHQ team to discuss how we can support you.
