If the Australian Government does not implement effective and coordinated governance frameworks for artificial intelligence (AI) systems now, this technology will soon outpace its ability to do so. This report presents the findings of a parliamentary inquiry into AI use in the public service and recommends the urgent implementation of robust governance frameworks to effectively balance innovation with risk management.

The authors note that these frameworks must ensure any risks, including sovereign risks, and biases that result from the adoption of these technologies can be effectively mitigated. This includes inadequate datasets, biometric biases and inaccuracies, disinformation and propaganda, foreign and electoral interference, online harm, cyber-crime and copyright violations.

Key findings

• Existing AI governance and oversight may not match the technology's rapid advancement.
• There are risks related to transparency, bias, security, and unintended consequences with the use of AI.
• Many public sector staff do not feel adequately trained to use AI tools effectively.

Recommendations

1. The Australian Public Service Commission (APSC) should, as soon as possible, introduce questions on AI use and understanding into its annual APS Employee Census.
2. The Australian Government should, within 12 months, convene a working group to consider AI legislation, update the Archives Act 1983, establish mandatory rules and governance, develop a training framework, and increase international cooperation.
3. The Australian Government should establish a Joint Committee on Artificial Intelligence and Emerging Technologies to provide effective and continuous Parliamentary oversight of the adoption of these systems.
4. Any guidance issued by the Digital Transformation Agency (DTA) or other agencies should clearly define all AI systems and applications and develop separate guidance for each.