That sequencing is the whole problem. The safe harbour is assessed retrospectively against the materials the director actually engaged with. If the understanding of the law that informed the decision was wrong, the documented record now works against the director rather than for them. This is exactly the area where the failure mode of AI legal research stops being an inconvenience and becomes a liability.
The safe harbour operates as a complete defence to a claim that a director or officer breached the objective standard of care in s 180(1), which requires the degree of care and diligence a reasonable person would exercise in the same circumstances. Where s 180(2) is satisfied, a court cannot measure the director's conduct against that objective standard.
The protection goes no further. A director whose conduct also engages s 181 (good faith and proper purpose), s 182 (improper use of position), or s 183 (improper use of information) derives nothing from s 180(2) in respect of those provisions. A research tool that returns a confident answer to "does the business judgment rule protect this decision" without surfacing that the safe harbour is confined to s 180(1) has not given a partial answer. It has given a misleading one, and a director who acts on it has documented their reliance on a defence that never covered the exposure that matters.
The threshold question is whether the conduct constitutes a "business judgment" within the meaning of s 180(3): any decision to take or not take action in respect of a matter relevant to the business operations of the corporation. Omissions fall within scope, which matters where the complaint is that a director failed to act. Purely administrative decisions that are not business operations decisions may not qualify, a boundary that has not been extensively litigated at appellate level and therefore one where retrieval alone will not give you a clean answer.
The safe harbour is lost on failure of any single condition, regardless of how clearly the remaining three are met. Identifying which element is vulnerable on the facts is where the defence analysis should begin, and three of the four turn on what the director did and documented at the time.
Good faith for a proper purpose. The director must have made the judgment in good faith for a proper purpose. Good faith requires an honest belief in the legitimacy of the decision; proper purpose requires that the judgment was made for purposes within the scope of the director's authority. The proper purpose test originates in Howard Smith Ltd v Ampol Petroleum Ltd [1974] AC 821 and asks whether the substantial purpose of the decision was within the range of purposes for which the power exists. These two elements are analytically distinct. A judgment made in honest belief but for an improper purpose fails this limb. A decision to dilute a shareholder's stake as the dominant objective, rather than as an incidental consequence of legitimate capital raising, illustrates the kind of purpose that will not satisfy the requirement.
No material personal interest. The director must not have had a material personal interest in the subject matter of the judgment. An interest is material where it may reasonably be expected to influence the director. A director who holds a material personal interest and neither discloses it nor abstains loses the safe harbour regardless of the quality of the decision itself. Where related-party arrangements are involved, the interaction with the financial benefits regime in Chapter 2E and the approval requirements under s 208 warrants concurrent assessment, and this is precisely the kind of cross-provision interaction a single-answer tool tends to flatten.
Informed to the extent reasonably believed appropriate. This is the element most commonly underestimated and the one where AI-assisted research most directly shapes the outcome. The director must have informed themselves about the subject matter to the extent they reasonably believed appropriate. The language is calibrated: it does not demand exhaustive investigation, but it requires the director to have actually formed a reasonable belief about what level of inquiry was sufficient and to have acted on it. ASIC v Hellicar (2012) 247 CLR 345; [2012] HCA 17 addressed the standards applicable to directors approving corporate communications and reinforces that passive reliance on management without genuine engagement with the substance will not satisfy a reasonableness standard.
The parallel is uncomfortable and worth stating plainly. Approving a legal position on the strength of a summary you did not interrogate is the same posture Hellicar warns against, transposed from a board paper to a research output. If the summary cited a case that does not say what the summary claims, or relied on a version of the provision that has since been amended, the director's "reasonable belief about what inquiry was appropriate" rests on a foundation that examination will expose. The element is assessed against the documented process: the materials reviewed, the questions asked, the advice obtained. That record is what makes the element provable after the fact rather than merely asserted, and it is fixed long before proceedings begin.
Rational belief that the judgment is in the corporation's best interests. The director must have rationally believed the judgment served the corporation's best interests. Section 180(2) provides that a belief is rational unless it is one no reasonable person in the director's position would hold, combining a subjective and an objective element. A judgment that destroys corporate value without commercial rationale, or that reflects indifference to the corporation's interests, will not attract the protection.
Section 187 provides a specific modification for directors of wholly-owned subsidiaries: a director may be taken to act in good faith in the best interests of the subsidiary where the subsidiary's constitution expressly permits it, the holding entity is not insolvent, and the director acts in good faith in the best interests of the holding entity. Confirm the constitution contains the s 187 authorisation before the holding company variant is relied upon, because a tool that does not have the constitution in front of it cannot tell you whether this pathway is even open.
The case law on the "informed themselves" element and the rational belief standard is voluminous and not consistently indexed by concept. Gathering how courts have treated the adequacy of inquiry across contested matters, tracing which factual configurations have defeated the safe harbour, and mapping the interaction between s 180(2) and the duties it does not cover is extraction and synthesis work that used to take hours. Done against authoritative Australian primary sources, it is the scaffolding that lets you spend your time on the judgment call rather than the retrieval.
What no tool can do is tell you whether this director's belief was rational on these facts, or whether the inquiry they made was reasonable in their position. That is the evaluative exercise the statute reserves for the court and the advice the practitioner is actually paid for. The danger is not that AI replaces that judgment. It is that a confident, well-organised summary feels like the answer when it is organised input to an answer that still requires you, and that the director's record will later be read as if they treated it as the answer.
The four conditions in s 180(2) are conjunctive: the safe harbour is lost if any single element is not satisfied, regardless of compliance with the others.
The "reasonably believed appropriate" standard under s 180(2)(c) requires genuine engagement with the subject matter; documented evidence of that engagement, including advice received and the authority it rested on, is what makes the element defensible when the decision is later scrutinised, so verify the law your advice relies on before it goes into the record.
The safe harbour addresses s 180(1) liability only; concurrent exposure under ss 181, 182, or 183 is not covered, and any research that suggests otherwise should be checked against the provisions directly.
Confirm the conduct falls within the s 180(3) definition of "business judgment" before relying on the safe harbour.
For directors of wholly-owned subsidiaries, verify the constitution contains the express authorisation required for s 187 to operate.
Because the "informed themselves" element is judged on the record built at decision time, practitioners working through s 180(2) defences can pull the current line of authority in Habeas, with results cited to paragraph level and the underlying extracts available for review, so the law the advice rests on is confirmed against the source before it becomes the record a court will examine.
Hero image: Ojus Jaiswal on Unsplash
