A review of evidence submitted to the Australian Parliament’s Joint Select Committee on Artificial Intelligence has found broad support for AI among contributors, but ongoing uncertainty over accountability when AI agents act on behalf of consumers — including in cases where an agent operates within authorised permissions but beyond what a user intended.
The findings are contained in a 106-page report, AI on Australia’s Terms: What Parliament’s AI Inquiry Reveals About Power, Value and Accountability, published by the Sydney-based Humanware Institute, with Kate Cooper listed as principal author. The report analyses 564 public submissions (6,158 pages), ten hearing days held between 18 September and 22 October (53.4 hours), and 66 answers to questions on notice published by 28 October, representing 575 distinct contributors.
According to the report, nine financial-system institutions — described as three major banks, their peak body, four regulators and Treasury — gave evidence to the inquiry. It says that when asked who is responsible when an agent acts for a consumer, “none said who carries the loss” in the scenario where an agent acts within granted permissions but not in line with the consumer’s intent. The report argues that existing law applies to AI, but that how it applies to that case remains contested or untested.
The report also points to agentic payments already occurring in Australia, citing evidence that Mastercard reported Australia’s first authenticated agent transactions in January 2026 in trials on two major banks’ cards. It notes that the Reserve Bank’s Review of Payments System Regulation, which it says covers agentic commerce, is due to announce conclusions in 2028.
Cooper said the inquiry record showed contributors generally supported AI but sought conditions on its use. “The people and organisations who wrote to Parliament mostly said yes to AI. Then they set terms,” she said, including skills, consultation, review rights, and accountability when AI acts on people’s behalf.
Among the report’s headline figures, it says 367 of 575 contributors supported AI, and 290 of those attached conditions. It says organisations were more supportive than individuals, reporting that 109 of 207 individual contributors were cautious or opposed.
The report identifies AI skills and literacy as the most common request, saying 138 contributors raised it. It also states that 82 contributors asked for notice, reasons and a human review that can change an automated decision, while 61 asked for clear liability across the AI value chain.
On workplace impacts, the report says 17 of 18 unions asked for consultation before AI changes work, compared with one of 35 banks and other businesses. It cites evidence from the Finance Sector Union that only two employers in recent years had been prepared to consider an AI consultation clause.
The report also raises rights and enforcement issues in creative work, citing a submission from a voice actor who said he found his cloned voice earning money on a major video platform and was unable to find an effective path to enforce his rights. The report notes that “a voice is not a copyright work”, and cites evidence from the ABC that it has not been paid by any frontier AI developer for training use.
On sovereignty and resilience, the report argues that domestic hosting alone does not guarantee public benefit, citing evidence including a statement from UNSW that future capacity for planned AI compute was already fully reserved and “most tenants are offshore companies”. It also describes evidence from two consultancies about a 13 June 2026 US Department of Commerce order that temporarily suspended access to two Anthropic models for users outside the United States, with access returning on 1 July.
The Humanware Institute recommends addressing responsibility for AI agents in payments as an early priority, describing the consequences as “tangible” and noting that rules already exist for unauthorised transactions, while the agent-permission-versus-intent scenario remains unsettled.
Separately, the report maps responsibilities across at least 19 Commonwealth ministers and agencies, alongside Parliament and the states. It calls for a single named point of accountability for coordination, pace and reporting across existing institutions rather than establishing a new regulator, alongside properly resourced regulators and “credible independent assurance”.
The committee’s final report is due by 30 November. The Humanware Institute says it will assess the committee’s final report against seven questions set out in its analysis, and that its report will be updated weekly as additional inquiry material is published.
