A Trust Framework for Government Use of Artificial Intelligence and Automated Decision Making

Public trust in government AI requires more than efficiency gains. The paper argues that speed and cost-effectiveness alone are insufficient to justify automated decision-making systems in the public sector. Trust must be actively earned through deliberate frameworks, particularly as the "trust bubble" created during COVID has already dissipated in many countries.

Government AI systems face unique accountability challenges under Administrative Law. Unlike private sector applications, public sector AI and automated decision-making must meet specific legal and constitutional responsibilities to citizens. The framework addresses how to ensure these institutions remain accountable to the public while adopting automation technologies.

A practical, jurisdiction-specific framework is needed alongside principles-based approaches. While principles-based frameworks (like Australia's AI Ethics Framework) provide important guidance, the paper proposes that governments need concrete, implementable frameworks tailored to their specific legal and institutional contexts to actually ensure ethical AI deployment.

The framework aims to assure positive human and societal impact, not just compliance. The paper emphasizes that responsible government AI must be designed to deliver genuine benefits to people and society, with explicit consideration of human and ethical outcomes as core objectives rather than afterthoughts.

This approach is designed for public sector context but has broader applicability. Though developed specifically for Australia's constitutional and jurisdictional environment, the authors indicate the framework is extensible to other countries and potentially to private sector applications.