20 Australian government departments use AI, but no one is watching closely
AI is used across 20 federal departments, but no one is checking for unapproved, 'shadow' use.
Twenty federal departments are running hundreds of AI experiments. Tools draft documents, summarise files and analyse workflows.
Shadow AI use by staff on personal accounts is what experts worry about most. Auditors found unapproved AI in the federal health department.
It had not yet assessed the risks. Health has 46 new AI projects, including one that summarises work plans.
But it relies on tired staff to click links and spot AI errors. The Department of Industry, Science and Resources says its AI trials are tested with humans checking for bias and errors.
Law professor Kimberlee Weatherall says the real problem is that departments assess themselves. The Digital Transformation Agency cannot enforce standards.
"There's no real reporting, and agencies decide what they tell us," she said.
- 20 federal
- Departments using AI
- 46 new
- Health department AI projects
- Shadow AI use on personal accounts
- Key concern
Why it mattersIf government decisions are based on AI summaries with hidden errors, and no authority is catching it, citizens have no assurance decisions are sound.
AustraliaAustralian government services rely on AI tools with gaps in oversight, raising questions about whether decisions affecting public services are safe.
✓ Claims checked against the source. checked 2 h ago
Open this story in InSnip →
