Governance in the Age of Artificial Intelligence: Key Takeaways from the 2026 Harold Ford Memorial Lecture

The integration of Artificial Intelligence (AI) into commercial enterprises has shifted from an aspirational technology initiative to an immediate corporate governance reality. This reality was underscored by the Hon. Andrew Bell AC, Chief Justice of New South Wales, in his delivery of the prestigious 2026 Harold Ford Memorial Lecture – Governance in the Age of Artificial Intelligence.

In his address, the Chief Justice provided a definitive signal to corporate Australia: AI is not a technical issue to be isolated and delegated to senior management, but a core governance matter that sits squarely at the board table. As regulators like ASIC and APRA accelerate enforcement actions regarding inadequate technological oversight, directors must urgently evaluate how their boards interact with automated tools.

The Onerous Reality of “Technology Neutral” Duties

A pivotal theme of the lecture is the Commonwealth Government’s technology-neutral approach to corporate regulation. Australia is not introducing AI-specific modifications to the Corporations Act 2001 (Cth); instead, the existing statutory framework applies directly to every decision touched by machine learning. Under Section 180(1) of the Act, directors must exercise the degree of care and diligence that a reasonable person would in their position.

If a company deploys an AI system that introduces unlawful bias, hallucinations, or data privacy breaches, the judiciary will critically examine what the board knew or ought reasonably to have known. Directors face a complex dual responsibility: while uncritical reliance on AI can lead to a breach of duties, a growing body of legal scholarship suggests that failing to use available AI tools to adequately interrogate data may also constitute a failure to remain properly informed.

Navigating the Safe Harbour “Accountability Gap”

The Chief Justice highlighted a profound “accountability gap” when applying existing statutory safe harbours to AI-assisted decision-making. Statutory provisions under the Corporations Act are structurally designed around human-to-human relationships:

  • The Problem with Delegation: Section 198D allows boards to delegate powers to a “person,” meaning a direct delegation of authority to an automated model receives zero statutory protection.
  • The Deficit of Independent Reliance: Section 189 protects reliance on expert advice, but explicitly requires that such reliance be underpinned by an “independent assessment” of the information. The opaque, “black box” nature of complex algorithms—where training parameters and reasoning pathways are hidden—renders true independent assessment exceedingly difficult, if not impossible.
  • The Inapplicability of the Business Judgment Rule: Section 180(2) demands a conscious, rational human reasoning process. The Chief Justice noted it is “extremely doubtful” that a director who blindly adopts an AI-generated recommendation can avail themselves of this defence.

The Psychology of the Boardroom and “Shadow AI” Risks

Beyond the letter of the law, the lecture warned against the insidious erosion of healthy courtroom and boardroom dynamics. Chief Justice Bell acutely observed that generative AI produces outputs with a level of “great confidence and clarity of language” that is shared with “the most accomplished of fraudsters”. This authoritative presentation risks inducing cognitive surrender, suppressing the productive conflict, robust debate, and diverse perspectives essential to proper oversight.

Compounding this cognitive risk is the material hazard of “shadow AI” use. When organisations fail to provide secure, sanctioned tools, personnel routinely turn to public consumer APIs to summarise documentation or draft records. This introduces severe, unmonitored liabilities regarding commercial confidentiality, cybersecurity, and the permanent loss of legal professional privilege.

Athena Board: Practical Governance by Design

To safely cross the chasm between blind reliance and wilful avoidance, boards require custom, enterprise-grade architecture. The Athena Board governance platform serves as a benchmark for how to implement AI responsibly and within the bounds of judicial expectations.

Athena Board resolves the critical deficiencies highlighted by Chief Justice Bell through a strict framework of data protection and augmented oversight:

  1. Zero-Training Isolation Boundaries: To address the profound privacy risks of public models, Athena Board guarantees that customer data stays entirely within its secure ecosystem. Sensitive corporate information, pre-release financial metrics, and legal briefings are strictly precluded from training third-party or public AI models, fully neutralising the risk of intellectual property leakage.
  2. Combating Cognitive Deference: Rather than presenting an opaque, definitive conclusion that invites groupthink, Athena Board’s engine acts as an analytical tool. It processes dense, multi-hundred-page board packages to explicitly parse apart core strategic claims, underlying risk drivers, operational anomalies, and compliance benchmarks.
  3. Facilitating Independent Assessment: By translating unstructured data into a transparent, targeted director inquiry trail, the platform equips directors with specific lines of questioning. This actively encourages the “crucible of debate” demanded by the Chief Justice, ensuring that the human director remains the ultimate, informed anchor of independent reasoning and corporate accountability.

The space between automatic deference to automation and stubborn resistance to progress is where modern governance must operate. By adopting secure, structured ecosystems like Athena Board, corporate directors can confidently fulfill their fiduciary obligations, maintaining absolute control and compliance in an AI-accelerated commercial landscape.

Athena Board can help, contact us at sales@athenaboard.com.