Australian AI governance overhaul proposed in Bond Law Review

By Marcus Bennett

A Bond Law Review paper reported on 10 November 2026 calls for Australia to adopt broad artificial intelligence legislation, mandatory board oversight and independent certification before advanced AI systems become more deeply embedded in organisations. The paper, Artificial Intelligence and the Law, also seeks stronger corporate governance rules as AI adoption expands across business and professional services.

Dr David Millhouse, Honorary Adjunct Senior Research Fellow at Bond University's Faculty of Law, says existing Australian law cannot keep pace with that adoption. His warning targets directors, trustees, legal practitioners and regulators, who face growing legal exposure as AI tools spread through workplace processes and decision-making systems.

“AI governance is not optional and not delegable,” Dr Millhouse said.

Millhouse's paper sets out recommendations, not enacted law. Its central policy claim is that Australia continues to avoid AI-specific legislation under a government policy still under review, while existing laws fail to keep pace with AI adoption.

What legal gap does the paper identify?

Millhouse's study examines legal and regulatory issues in three areas: corporate governance, legal practice and business operations. It identifies six risk categories extending beyond model errors to data ownership, privacy, professional duties and automated accountability.

• The paper flags inaccurate or biased outputs.
• It identifies uncertainty over data ownership.
• It lists privacy breaches.
• It lists intellectual property disputes.
• It raises professional negligence as a legal risk.
• It questions accountability in automated decision-making.

Millhouse links those six risk categories to board oversight as a liability issue rather than a voluntary technical exercise. His paper places responsibility for AI supervision inside governance structures, not solely with information technology teams.

Why does Australia need AI-specific legislation?

Millhouse says Australia should move beyond its current approach of avoiding AI-specific legislation. He describes the government's position as a policy error and calls for broad reform before more advanced AI becomes widely embedded in organisations.

“Present Australian government policy (under review) is to not legislate a specific AI statute. This is an error requiring legislative rectification in the pre-AGI era. Reform needs to be capacious,” Dr Millhouse said.

He frames the legislative window as the pre-artificial general intelligence (AGI) era. The paper describes a broad statute addressing AI's legal effects across corporate governance, legal practice and business operations.

What would a national framework require?

The paper sets out governance measures for boards, regulators and professional bodies. Five core proposals cover institutional oversight, workforce preparation, professional disclosure, external assurance and company law.

• Millhouse proposes expanding the role of the Australian AI Safety Institute.
• He calls for mandatory governance training.
• He seeks tighter disclosure obligations for professional use of AI.
• The paper proposes independent certification and auditing arrangements.
• It calls for changes to the Corporations Act 2001, including stronger protections for data ownership and intellectual property.

How would boards change?

At company level, Millhouse says boards should create AI-specific oversight structures rather than leave supervision solely to IT teams. Another recommendation would give each board a Digital Readiness Officer and an AI compliance committee with standing similar to an audit committee.

Millhouse proposes that those bodies monitor AI implementation, legal risk and arrangements for human supervision. His governance model would keep AI compliance separate from operational IT management.

“Boards should appoint a Digital Readiness Officer to implement IEWA [Intelligent Enterprise Workflow Automation] reporting through an AI compliance committee equivalent to the audit committee and separate from operational IT management,” Dr Millhouse said.

The paper also directs listed companies and trusts to run board-level workshops on internal AI policies and compliance reporting. Management would then receive regular training, with responsibility for assessing system behaviour and supervision kept clearly inside the governance structure.

What would change for regulated professions?

Millhouse's professional-liability proposal combines an outcomes regime with certification and disclosure. He would require lawyers and barristers to disclose their deployment of AI and the extent of professional human supervision.

“Implement a bias and factual data and outcome regime policed by an independent certification authority (eg AAISI, if its mandate is reformed); introduce lawyer and barrister mandatory certification disclosing deployment of AI and the extent of professional human supervision,” Dr Millhouse said.

He gives AAISI, the Australian AI Safety Institute, as the example certification authority if its mandate changes. The paper also aims its governance proposals at regulators and professional bodies, placing them alongside boards as participants in compliance.

Can AI improve legal work?

Millhouse's research extends beyond compliance to legal linguistics, the study of how law and language interact. He considers whether combining AI with that field could improve legal accuracy across languages and cultures.

The paper places this possibility within a wider debate about generative AI's effect on legal work. Millhouse's paper says adoption is increasing and expectations of productivity gains remain high, yet the scale and nature of the change remain uncertain.

“It is too early to definitively predict how generative AI will affect the law and the legal profession. How much of the hype will come to pass, and whether the effects of AI will be revolutionary or evolutionary is not yet obvious. Technology may be revolutionary, but its application is evolutionary,” Dr Millhouse said.

Millhouse's formulation separates technical capability from institutional adoption. Generative AI may produce a large technical shift, while its application in legal services changes gradually through governance, training and review.

Who is behind the proposal?

Millhouse serves as Honorary Adjunct Senior Research Fellow at Bond University's Faculty of Law. He has previously written on corporate governance, law reform and financial regulation.

He also founded and chairs ActiumAI, an Australian company focused on intelligent enterprise workflow automation and corporate governance. That focus connects directly to the paper's IEWA reporting proposal, which would make intelligent workflow automation a board-level compliance subject.

What must happen before AGI?

The paper closes with a warning that the legal system and corporate oversight structures should adapt before AI becomes deeply embedded in critical decisions. Millhouse places the reform agenda in the pre-AGI period rather than after AGI arrives.

The paper identifies the Australian government, regulators, company boards, trusts and professional bodies as the institutions responsible for the resulting governance tasks. Its practical deadline is the point at which critical decisions run through AI without matching legal accountability.

“It would be trite to forecast epochal change, but to emulate Sir Winston Churchill, the journey to artificial general intelligence (AGI) is at 'the end of the beginning',” Dr Millhouse said.