Australia’s shift to digital government services has resulted in platforms like MyGov, which one million Australians visit every day. It comes as no surprise then that in the 12th biennial survey on e-government by the United Nations*, Australia ranks 7th among the leading countries in e-government development. However, this achievement also means citizens expect a seamless digital experience while expecting higher standards of governance and fairness than from the private sector. Digital services are facilitated by data resources and their analyses, including methods that utilise artificial intelligence.
Public trust in the use of assets and data resources needs to be high, and a recent whitepaper* posits that public citizen services are held to a higher account than private companies. Therefore, AI adoption in the public sector must navigate specific barriers to achieve world-class digital government while maintaining the trust of its citizens.
AI governance is a constant and critical consideration
Governance is a continuous spectrum, from the administration of citizen and government data through to the governance of AI itself. Many AI projects start off in testing and development environments, but without the right guardrails in place, they do not move into production. Strong and clear data and AI governance will produce two outcomes:
- The ability to inspect work programs according to a set of principles is crucial.
- The system continuously provides detailed information about models in their environments.
This builds a foundation of trust, enabling users and citizens alike to benefit from the advances of AI.
Simpler, robust processes are imperative to maintain trust in government process
However, with the rise of new technologies and speed to market outweighing proper processes, trust in government is being eroded across OECD countries. Citizens can experience complicated application processes and long wait times when applying for services. This typically happens after a life-changing or significant traumatic event, when difficult-to-navigate processes are already challenging.
In discussions with public service leaders and staff, we see strong support for testing out new applications of AI, particularly generative AI, where deep learning in a dynamic environment allows for logical deduction and pattern identification. This in turn addresses the need for fast and secure e-government services via customer-facing applications, as well as predictive modelling for better future outcomes in policy. Technology improvements have broadened the applicability of AI models, but there remains a large degree of caution in applying these models at scale.
Simply put, these models cannot exist without the right governance structures in place.
In a recent white paper, ‘Pathways to Trusted Progress with Artificial Intelligence’, leaders across the public sector with roles in digital government were brought together to discuss adoption and strategies to enhance trust in the use of AI in government.
Read here as esteemed authors Kevin Desouza and Gregory Dawson explore the principles of explainability, transparency, and stability and how this relates to public trust.