The Role of AI in Reducing Regulatory Costs
When it comes to building or making anything, there’s usually paperwork involved. Now, politicians have a new favourite weapon in their age-old fight against red tape: artificial intelligence (AI). This year’s federal budget was filled to the brim with examples of ways the government hopes to use the new technology to save $10.2 billion in regulatory costs. Using AI to speed up medical and environmental approvals were two headline examples.
Amid the promises of efficiencies, the budget contained only a birds-eye view of the exact role AI will play in these government processes and the safeguards around them. How will AI be used to cut red tape? The idea of handing over part of the process of approving medicines or housing plans to the machines might prompt fears about whether we can trust them with that responsibility.
Based on what we know, the government isn’t delegating the decision-making to AI. Rather, it’s being used to help with the paperwork around those decisions. The medicine approval process by Australia’s medicines regulator, the Therapeutic Goods Administration (TGA), already considers other countries’ approvals of medicines as a shortcut to avoid drug makers repeating unnecessary trials. This all goes into the final decision by Australian regulators about whether a medicine will be safe for Australians.
According to this year’s budget, the TGA is “using AI to evaluate the suitability of medicines already approved by another comparable overseas regulator”. In this case, AI will compare other countries’ regulations and, the government claims, it will save manufacturers $340 million each year.
International Examples of AI in Regulation
Have any other countries done the same? The UK had some success using a similar approach. In October 2025, its Medicines and Healthcare Products Regulatory Agency said clinical trial approval times fell from 91 days to 41. The UK process involved clinical trial approval, not the marketed-medicines reliance pathway the TGA is automating, but the structure — AI assists, humans decide — is the same.
What about AI and housing approvals? The other big item flagged in the budget was building an AI tool to help housing developers through the environmental approval processes, as part of a $105.9 million package to modernise data and systems belonging to the Department of Climate Change, Energy, the Environment and Water and the National Environmental Protection Agency.
This will be specifically used to “assist and better inform proponents seeking environmental approvals”. A spokesperson for the Department of Climate Change, Energy, the Environment and Water said it had already begun testing a proof-of-concept AI tool. AI will be used to “enhance the user experience for people navigating complex laws and documentation” when they go through the approval process but leaving the decision-making to staff.
“Decisions about whether to approve projects must, and will, always be made by assessment officers, not by AI.” The aim is to get “faster yeses and faster noes”, as described by minister Murray Watt last year when he spruiked using AI in the housing environmental approval process.
Other AI Projects in the Budget
What are some of the other ways AI was spruiked in the budget? A list of similar projects promised or promoted in the budget included:
- The Department of Veterans’ Affairs trialling an optional tool to pull information out of large claims documents to cut manual review time
- IP Australia’s use of an AI-assisted trademark availability tool
- IP First Response, a chatbot for small businesses with infringement queries
- The National Library of Australia transcribing 58,000 hours of oral history using AI
Expert Opinions on AI in Government Processes
What do the experts say? Queensland University of Technology law professor Nic Suzor, who has spent 20 years researching government information access, said this is where current-generation AI is most useful. “This is both inevitable and mostly good — [it] always depends on the details,” he said. “But doing that integration work is exactly where you can get really big benefits from current-gen AI models.”
Professor Suzor said the greatest benefit and least risk came from leaning on AI to open up and index data scattered across departments, a formidable issue for government and other big organisations. He said there were always ethical considerations like privacy when it came to doing this.
Australian National University’s Responsible Innovation Lab’s senior lecturer Ehsan Nabavi wrote in the Conversation last year about his broader concerns about AI introducing biases into housing approvals, a process proposed and piloted by some of the states. “[AI] shapes what gets seen and what gets ignored in different stages of assessment, often in ways that aren’t obvious at all,” Nabavi wrote. He also warned that viewing AI as a quick fix to some of our thorniest problems was likely fool’s gold. “It risks distracting deeper systemic issues such as labour market bottlenecks, financial and tax incentives, and shrinking social and affordable housing.”
