HTI welcomes the Albanese Government’s commitment to address significant policy failures found by Robodebt Royal Commission
Lessons for government in use of technology for service delivery
‘Robodebt’ operated between 2015-2019 as a scheme to recover overpayments made to Australians in the social security system, using a largely automated calculation system and issuing debt notices to welfare recipients.
However, the scheme produced unacceptably-high rates of error, and was ultimately found to be unlawful – resulting in Robodebt payments by individuals being repaid by the Government.
The unlawful operation of this debt-recovery scheme had a devastating impact on the lives of many Australians, including the most vulnerable in our community. Acting on the Royal Commission’s findings will help bolster trust in government, particularly where new and emerging technologies like AI being used to deliver social security services at scale.
The Royal Commission has made several recommendations to address the harms that can occur when government uses technology without adequate legal and operational safeguards. Given the rapid uptake of AI by government to make administrative decisions and deliver essential services, urgent reform is needed.
In its response to the Royal Commission, the Albanese Government has accepted key recommendations that will go some way to ensuring accountability for the use of automated decision-making tools and AI systems by government, including:
- establishing a legal framework for the use of automation in government services; and
- empowering a new or existing body to monitor and audit automated decision-making processes with regard to their technical aspects and their impact in respect of fairness, avoiding bias, and client usability.
Beyond Robodebt, the Royal Commission's findings hold valuable lessons to guide government use of technology for service delivery into the future. Building community trust in the use of new technologies like AI will require regulation that is underpinned by:
Transparency. Citizens should be informed that AI is being used to make a decision that will impact them, and be provided with enough information to challenge that decision and seek redress. The automated ‘Robodebt’ letters sent to individuals provided no explanation for the basis of debts raised, meaning they could not assess if the debt was incorrect or unlawful, and therefore could not challenge it without great difficulty.
Accountability. Robodebt continued unabated due to failures of both external and internal accountability and the absence of oversight mechanisms. Digital government service delivery should comply with relevant law, be consistently and continuously monitored, and ensure there are clear redress mechanisms in place for affected individuals.
Human rights and human values. A major theme of the Royal Commission report was the lack of care and consideration shown to the people affected by the scheme. Before new technologies are introduced, it is essential to assess the impacts on people, in consultation with those affected, and reflect these considerations in the design and implementation process.
The Robodebt Royal Commission findings should be a watershed moment for the use of new and emerging technologies by government. Specifically, the findings should be located within a co-ordinated AI regulatory reform agenda that will address the complex range of issues and applications associated with AI across different sectors and areas of law.
In 2024, in the context of its Future of AI Regulation Program, HTI will continue widely consulting with stakeholders to develop a cohesive roadmap for regulation that responds to the risks and benefits posed by AI, centred on human impacts, aligning with human values and protecting human rights.