The Australian Taxation Office is building and deploying AI models in-house without centralised policies and guidance for public servants, exposing it to legal and ethical risks, the auditor-general has warned.
Human oversight has also been found lacking, with the agency not adequately monitoring its 14 newest AI models, some of which automatically flag potential non-compliance in tax affairs for review.
It means that one of the government’s heaviest users of AI cannot be certain its models are free from bias or safe, the Australian National Audit Office (ANAO) said in a report labelling the tax agency’s AI governance only “partly effective”.
Want to know how this story ends?
Become a subscriber for complete access to all stories on InnovationAus.com and our daily newsletter.
Do you know more? Contact James Riley via Email.