Artificial intelligence is pushing false information into Australia’s policymaking process, and parliamentary committees are now being forced to question the evidence arriving at their doors. Inquiries are receiving submissions that invent studies, cite nonexistent research and attach false claims to real academics and authors.
The warning is direct: material that looks researched may not exist at all. As AI-generated content spreads through public submissions, committee members face a new challenge—separating genuine evidence from convincing hallucinations before reports shape policy.
Invented research reaches parliamentary inquiries
Parliamentary committee chairs have warned that AI-generated material is introducing “hallucinated” information into policymaking. Hallucinations happen when large language models produce or refer to content that appears real but does not actually exist.
That problem is now showing up inside formal inquiries. Committees have been flooded with content created by AI, including submissions that invent studies and attribute nonexistent research to real academics and authors. In some committee reports, the majority of cited sources in a submission appear to be AI-generated hallucinations.
The danger grows when false references sit beside genuine evidence. A submission can look serious, use the names of real researchers and present claims in a polished format, yet still lead readers toward information with no real source behind it.
Paul Scarr, a Liberal senator, said the reporting showed clear dangers in the rising number of submissions containing hallucinated content. He identified the central risk as a failure of decision-making, whether the decisions happen in parliament, business or another setting.
“The biggest risk is that decision-makers, whether or not they’re in parliament or in business or anywhere else, make decisions based on incorrect information,” Scarr said.
Committees face a new evidence-checking burden
Scarr said committees must examine submissions with greater care, checking footnotes, sources and claims instead of treating polished writing as proof. The new pressure reaches beyond identifying obvious errors because invented material can hide among legitimate arguments and real-world evidence.
“We’ve been put on notice. We really need to check and go the extra mile in terms of checking footnotes and sources and claims which are made in submissions,” Scarr said.
He also mentioned a submission that included fake allegations of wrongdoing by KPMG, which the consultancy firm complained about. The case shows how hallucinated information can create consequences for named organisations as well as for the committees reviewing the material.
Louise Miller-Frost, a Labor MP and chair of the social policy and legal affairs committee, said committees are becoming more aware of the issue. Committee secretariats now recognise that some references included in submissions do not exist, making source checks a vital part of the inquiry process.
Miller-Frost described the kind of claim that should trigger a closer examination. “I think if they put in something unusual or really salient then you’d have to go down the rabbit hole and go, ‘OK I need to find out more information on that particular submission and then go into the references,’ and then it would be obvious that their references don’t exist,” she said.
AI submissions still have a place in public debate
The response from committees is not a call to discard every submission created with AI. Miller-Frost said AI-generated submissions should not be rejected simply because they used the technology, since their findings were not different from other evidence.
That distinction matters because parliamentary inquiries do not receive material only from organisations with large teams and extensive resources. They also hear from people who may struggle to prepare a submission without help.
“Some of [the submissions] are from agencies and academics and government departments and they’ve got lots of resources. But we also get, particularly in the social services area, some people who are individuals in the community, and they may not have the time and the capacity and availability to write what they want to write,” Miller-Frost said.
For those individuals, AI could help turn their views into a submission and give them a way to participate in policymaking. The challenge is making sure that assistance does not introduce invented studies, false references or allegations that were never supported by evidence.
- Committees are receiving AI-created submissions containing hallucinated information.
- Some submissions invent studies or attribute nonexistent research to real academics and authors.
- Committee reports have identified submissions where most sources appear to be AI-generated hallucinations.
- Committee secretariats are becoming more aware that some references do not exist.
- AI may help people who lack the time or capacity to write submissions, but claims still require checks.
Australia’s parliamentary committees now have to balance access with accuracy. AI can help more members of the public enter the policy conversation, yet every unusual claim, footnote and source must survive closer inspection before it influences a decision.
The next stage of this debate will focus on trust: not whether AI appears in a submission, but whether the evidence behind that submission can be verified.
Based on
