Justice Bell said relying on safe harbours built into the Corporations Act will only apply if directors can properly justify why they trusted delegates to use AI, in the 2026 Harold Ford Memorial Lecture at the University of Melbourne Law School on May 21. And he was particularly harsh on directors relying on the recommendations of AI bots, adding it would be "extremely doubtful" if they could use the statute's business judgment defence.
The judge suggested boards would be "prudent" to put in place policies which ensure management tells directors if they have used AI for reports, advice and recommendations, what was involved in vetting that AI, and whether the AI-sourced information is considered reliable and accurate.
"Directors must be vigilant and disciplined in their use of AI," Justice Bell said. "Apart from the confidentiality, data governance and proprietary concerns, none of which is insignificant, independence lies at the heart of good corporate governance, and great care must be taken by directors to ensure that the crucible of debate and productive conflict of ideas in the board room is not undermined or supplanted by AI."
The explosion in the use of generative AI since the introduction of ChatGPT roughly four years ago has been a "technological tsunami," and although AI-specific regulations lag behind the pace of technological change, there are a number of existing statutory and common law obligations that directors should heed in relation to AI, according to the judge.
Bell highlighted the March Federal Court decision by Justice Michael Lee in ASIC v. Bekier, a case brought by the regulator against the directors and CEO And Chief Legal Officer of Star Entertainment Group Ltd over alleged governance failures.
In that decision, Justice Lee said a required legal duty of all directors was to "take reasonable steps to place themselves in a position to guide and monitor the management of the company" and that they couldn't just rely on the advice of management in place of examining "an important matter that falls specifically with the board's responsibilities."
Given the quick and broad adoption of AI as a "critical component" of many companies' operations, it "might be argued" that the use of AI to help make decisions was that sort of important matter and that "close knowledge" of how a company uses AI to assist with decision making is part of directors' necessary "care and diligence," Justice Bell said.
Directors are typically responsible for the actions of their delegates, unless they believed "on reasonable grounds at all times," after a "proper inquiry," that the delegate was "reliable and competent" and would act in accordance with the duties imposed on the director.
That raises questions about how "reliability" and "competence" would be determined when a delegate uses AI, for example whether competent use of the technology to perform the required task is enough, or whether that would be considered a disallowed delegation of authority, according to Justice Bell.
And for the Corporations Act's "reasonable grounds" safe harbour to apply, directors still have to justify "why their trust in a delegate who themselves was relying on AI was reasonable, in light of what we now know about the limitations of AI, including its scope for seductive self-confidence and sycophantic agreement, bias, or hallucination," the judge said.
These concerns, among other AI-related issues, don't necessarily mean that directors need to avoid using AI, but they do indicate that directors should use the technology as a decision-making aid rather than a substitute for their own assessment, and maintain an ongoing awareness of AI's limitations and potential risks, according to the judge.
A representative for the Australian Chamber of Commerce and Industry, a business association that has previously commented on issues around the industry adoption of AI, did not respond to a request for comment on Tuesday.
For a reprint of this article, please contact reprints@law360.com.