What happened
Damian Warren was a long-serving supervisor in the Client Assets Branch (CAB) of NSW Trustee and Guardian, the agency responsible for managing the finances of people under financial management orders. In late 2011 the agency's CEO, Ms Imelda Dodds, commissioned an external consultant, Ms Joy Woodhouse, to conduct a "Workplace Review" after client complaints, an Ombudsman report and successive audit reports identified serious systemic failings. The terms of reference were deliberately broad: working arrangements, interpersonal relationships, culture, recruitment, delegations, structure, communication protocols and any other matters. Ms Woodhouse interviewed roughly two-thirds of CAB staff, including Mr Warren and the manager Ms Sharon Minns, on the explicit assurance that comments would be treated confidentially, no individuals would be identified in the final report, and the exercise was not a "witch hunt".
In April 2012 Ms Woodhouse delivered a 45-page Report marked "Confidential" together with a separate five-page Memorandum placed in a sealed envelope. The Memorandum recorded specific allegations of workplace misconduct made by staff about Mr Warren and another manager. These allegations were outside the original review scope but Ms Woodhouse felt duty-bound to pass them on. Ten days before any staff briefing, Mr Warren was suspended on full pay and given a letter from the Director-General of the Attorney General's Department that particularised the allegations against him and named the staff members said to be the victims of his conduct. The letter made clear the allegations had emerged during the Woodhouse review.
On 24 May 2012 Mr Warren lodged a GIPA request for the Report, its recommendations and Ms Woodhouse's notes, offering de-identification if necessary. The agency ultimately decided he could inspect the Report under supervision and take notes but could not have a copy. An internal review affirmed that stance. Only after Mr Warren commenced Tribunal proceedings did he learn of the existence of the Memorandum. On remittal under s 65 of the ADT Act the internal reviewer again refused a copy of the Report and decided the Memorandum was outside the original request. By the time of the Tribunal hearing the agency no longer pressed the "outside scope" argument but contended there was an overriding public interest against disclosure of the entire Memorandum. The disciplinary investigation was later completed and the allegations against Mr Warren were found unsubstantiated; he returned to work.
The Tribunal received the Report, the Memorandum and parts of Ms Dodds' evidence in confidence pursuant to s 107 of the GIPA Act. After a hearing on 21 February 2013, Deputy President Higgins delivered detailed reasons on 7 August 2013. She found the agency's decision-making approach fundamentally misconceived, set the decision aside under s 63(3)(d) of the ADT Act and remitted it for reconsideration on the basis of explicit findings about which categories of information must be released to Mr Warren by way of copy (with deletions) and which could be withheld.
Why the court decided this way
The Tribunal's reasoning rests on a close textual and structural reading of the GIPA Act. Deputy President Higgins began by emphasising the statutory objects in s 3 and the presumption in favour of disclosure in s 5. She reminded herself that the only public interest considerations against disclosure are those listed in the Table to s 14 or in Sch 1, and that the balancing exercise required by s 13 must be undertaken in accordance with the principles in s 15.
Crucially, she held that the structured decision-making process described by the Appeal Panel in Camilleri (at [25]) applies to the access decision under s 58, not to the subsequent administrative question of form under s 72. Once an agency has decided that access should be given, the default position is that it must be given in the form requested (s 72(2)). The four exceptions in s 72(2)(a)–(d) are narrow. Paragraph (a) concerns practical interference with the agency's processes of providing access, not its broader statutory functions. Paragraph (d) is engaged only where there is an overriding public interest against disclosure of the information in the way requested; it is not an invitation to re-run the entire s 13 test for every form-of-access dispute. This construction is reinforced by the long history of identical provisions in the former FOI Act 1989 and in other Australian jurisdictions, none of which had been read as allowing exempt-matter arguments to dictate form.
Applying that framework, the Tribunal accepted that cl 1(d) (prejudice supply of confidential information), cl 1(f) (prejudice effective exercise of functions) and cl 1(h) (prejudice effectiveness of a review) applied at an abstract level to the confidential staff contributions and to the description of Ms Woodhouse's methodology. Ms Dodds' evidence that staff had been promised confidentiality and that wholesale disclosure in a previous workplace had produced "catastrophic" distraction was accepted as establishing a reasonable expectation of prejudice. However, the Tribunal insisted on granular analysis. Information that was not confidential staff comment, that had already been disclosed in substance to staff, or that concerned only the terms of reference and high-level recommendations did not attract the same weight.
Personal information considerations under cl 3(a) had been overlooked by the agency except in relation to Mr Warren (where it was treated only as a factor in favour). The Tribunal held that cl 3(a) was a live consideration against disclosure for third-party staff information. Conversely, because the allegations against Mr Warren had already been particularised in the Director-General's suspension letter, that portion of the Memorandum had been "revealed" within the meaning of cl 1 of Sch 4. The cl 3(a) consideration therefore did not apply to it. The applicant's identity, his direct interest in responding to disciplinary action, and the principles of procedural fairness and natural justice were powerful personal factors in favour under s 55. Ms Minns had no objection to release of material concerning her. In contrast, the risk of "serious harassment or serious intimidation" required by cl 3(f) was not made out; subjective fears of some staff and general workplace tension fell short of the statutory threshold.
The result was a nuanced outcome: Mr Warren was entitled to a copy of the sections and attachments containing his own contributions, Ms Minns' contributions, and the specific allegations already revealed to him. All other matter—third-party personal information, unrevealed confidential allegations, and review methodology that could prejudice future exercises—could lawfully be deleted under s 74 before a copy was supplied. Because the agency had not approached its task on this basis, the decision was set aside and remitted with these binding findings.
Before and after state of the law
Prior to the GIPA Act the FOI Act 1989 contained a similar form-of-access provision (s 27(3)) but operated against a different public-interest backdrop. Exemptions were more numerous and included a broad "internal working documents" exemption. Agencies routinely refused copies of sensitive investigation reports on the basis that the document as a whole was exempt. The GIPA Act deliberately narrowed the considerations against disclosure to a closed list, abolished the conclusive exemptions (except Sch 1), and introduced an explicit objects clause and s 15 principles that forbid reliance on embarrassment or possible misinterpretation.
Warren is the first detailed Tribunal examination of how the new s 72(2)(d) interacts with the public-interest test. Before Warren some agencies and reviewers appeared to treat the form-of-access decision as simply another opportunity to re-weigh all s 14 considerations. The decision makes clear that this is wrong. The access decision comes first. Only after access has been granted (or partial access granted under s 74) does the agency ask whether the requested form would itself cause one of the four enumerated harms. The practical effect is to make unconditional copies the presumptive outcome wherever the access decision has favoured release, even of sensitive workplace review material.
After Warren the Information Commissioner issued guidelines that expressly cite the case for the proposition that form-of-access disputes should not become surrogate public-interest contests. Later decisions have applied the distinction between "access" and "form" without controversy. The case also sharpened the operation of cl 3(f): mere apprehension of workplace friction is insufficient; the risk must be of serious harassment or intimidation and must be objectively reasonable.
Key passages with plain-English translation
Paragraph [53]: "In my view, the respondent's approach is misconceived and contrary to the proper construction of s 72(2) and the decision making processes within the GIPA Act in regard to the decision about access and the decision about the form of access."
Plain English: The agency got the legal structure backwards. Deciding whether you get to see something is one question; deciding whether you get a photocopy or just a supervised look is a different, narrower question.
Paragraph [60]–[61] (adopting Camilleri): The Tribunal must first decide if a consideration against disclosure "could reasonably be expected to" have the prohibited effect at a general, abstract level before weighing that consideration against the specific factors favouring release in the instant case.
Plain English: You look at the type of information and the type of agency function first, without getting lost in the individual drama. Only after deciding the prejudice is real in a general sense do you weigh how strong the applicant's personal need is.
Paragraph [94]: "What must be established is a reasonable expectation of a risk of 'serious' harassment or 'serious' intimidation if the information is disclosed... Ms Dodds acknowledged that since his return to work, there was no evidence of the applicant having engaged in conduct of this nature, let alone, conduct of this nature which was serious."
Plain English: It is not enough that some staff feel nervous. The law demands proof of a real chance of serious bullying or threats. Past behaviour and the applicant's assurances mattered.
Paragraph [98]: "It is well accepted that information of this kind should be protected from disclosure: see Williams... the verbatim content of the interviews were not revealed in the investigation report."
Plain English: Even if you get the summary report, you do not automatically get the raw interview transcripts. That principle survives under the GIPA Act for material that has never been shown to the applicant.
What fact patterns trigger this precedent
Warren is triggered whenever an agency grants access (or partial access) to a workplace review, investigation report or similar document but then refuses the applicant's preferred form—usually a copy—on public-interest grounds. The paradigm case involves an employee who is the subject of adverse comment in a review that later spawns disciplinary action. Three elements must coincide: (1) the applicant seeks his or her own personal information or information already partially revealed; (2) procedural fairness or natural justice considerations are engaged; and (3) the agency has relied on cl 1(d), 1(f), 1(h) or 3(a) to justify withholding a copy.
The decision also applies to any attempt to re-litigate the entire public-interest test at the form-of-access stage, to any failure to differentiate between revealed and unrevealed personal information, and to any reliance on cl 3(f) without evidence of serious harassment risk. It is not limited to NSW Trustee and Guardian; any agency that commissions confidential reviews (health services, police, local councils, universities) must now approach form-of-access decisions on the narrowed basis set out in [53]–[59]. Requests for deletion under s 74 rather than outright refusal are the expected outcome.
How later courts have treated it
Warren has been favourably cited in numerous subsequent ADT and NCAT decisions. In Miskelly v NSW Department of Education [2014] NSWCATAD 26 the Tribunal applied the separation of access and form decisions to refuse an agency's attempt to impose a "view only" regime on a bullying investigation report; the applicant received a redacted copy. Ainsworth v NSW Police Force [2014] NSWCATAD 52 followed the abstract-prejudice analysis for cl 1(d) but distinguished Warren on the facts because the applicant was not the subject of the allegations and no disciplinary process was on foot.
The Appeal Panel in Department of Family and Community Services v Hall [2015] NSWCATAP 13 cited Warren with approval for the proposition that once information has been revealed in one document it loses cl 3(a) protection in related documents. More recently, Zonnevylle v Department of Education [2020] NSWCATAD 86 relied on Warren's granular balancing to order release of an applicant's own comments in a workplace culture survey while protecting third-party identities. No decision has doubted the central holding on the construction of s 72(2). The case is now standard citation in Information Commissioner guidelines on form of access and in agency training materials.
Still-open questions
Several questions remain unresolved. First, the precise boundary between "revealed" and "unrevealed" personal information is fact-sensitive. If an allegation is paraphrased in a suspension letter but the Memorandum contains additional verbatim detail or context, does the additional detail remain protected? Warren suggests it does, but later cases have not tested the outer limit.
Second, the interaction between s 72(2)(d) and s 74 deletions is not fully mapped. If an agency can delete exempt matter and still provide a copy, when (if ever) will provision of that redacted copy itself engage an overriding public interest against disclosure "in the way requested"? The Tribunal left open the possibility that widespread distribution risk could still justify view-only access in extreme cases, but gave no example.
Third, the weight to be given to an applicant's assurance that he or she will not harass or intimidate remains unclear. In Warren the absence of any post-return adverse conduct was decisive, but what if the applicant has a prior history? Clause 3(f) looks to future risk, yet agencies may be reluctant to rely solely on undertakings.
Fourth, the decision assumes the review was not itself an "investigation" for cl 1(h) purposes. Where a review morphs into a misconduct inquiry, the analysis may differ. Finally, the Tribunal did not decide whether an agency can impose a condition preventing further copying or distribution once a copy is released; s 73(2) permits conditions only to avoid an overriding public interest against disclosure, but the circumstances remain untested.
These open questions mean that while Warren provides a clear roadmap for the ordinary workplace-review case, agencies and applicants must still undertake careful, document-by-document analysis in borderline situations. Practitioners should pay close attention to the exact wording of suspension or show-cause letters, because that wording may determine whether material has been "revealed" and therefore loses its cl 3(a) protection.