Open government by design: lower the cost of transparency and rebuild trust

When openness is expensive, everyone loses

Australia’s FOI framework was created in a paper-records era. Public administration now produces email, chat, case-management records, data extracts, and electronic briefings at enormous scale, but access requests still trigger a largely manual sequence: interpret the request, search several systems, inspect every candidate record, consult internally, apply exemptions, redact, and issue a formal decision.

The Office of the Australian Information Commissioner (OAIC) reports that this process cost Commonwealth agencies and ministers $97.99 million in 2024–25. It consumed 1,171,890 staff hours—the equivalent of about 586 full-time staff years—and produced 25,211 decisions at an average cost of $3,887 each. Only 73% of requests were completed within the applicable statutory timeframe.

Charges do not solve that cost problem. Agencies collected $65,181 from applicants in 2024–25, equal to less than 0.1% of the system’s reported cost. Charging more may suppress requests, especially from freelance journalists and small community organisations, but it does not make a fragmented record system cheaper to search.

The Freedom of Information Amendment Bill 2025 proposed application fees for some requests, a ban on anonymous and pseudonymous requests, broader refusal powers, and a 40-hour processing cap. The bill was discharged from the Senate Notice Paper on 5 March 2026 and is no longer proceeding. Its approach still illustrates a recurring policy temptation: reduce the cost of transparency by making access harder rather than making disclosure easier.

Transparency is a service, not a compliance event

Trust is weakened when official information appears only after a long contest. Delayed releases are less useful to a journalist investigating a live decision, less useful to Parliament while legislation is being debated, and less useful to a citizen trying to understand an agency action that affects them.

The same process can be demoralising inside government. Information officers inherit broad requests because outsiders cannot see what records exist. Policy and legal teams repeatedly review similar material. Staff then experience scrutiny as a disruptive exception to their work rather than as a normal output of public administration.

A better system lowers the transaction cost on both sides. It gives the public enough information to ask precise questions, gives agencies structured records that are easier to assess, and reserves specialist judgement for genuinely sensitive material. That is how lower operating costs can support more—not less—accountability.

1. Publish a searchable register of public records

Norway’s eInnsyn service provides a practical model. Public bodies publish searchable journal entries for correspondence, records, meetings, and committee business. People can search anonymously and without charge, request a specific document, or read a document immediately when the full text has already been published.

Australia should create a common public record metadata standard and a national search service. Agency record systems would automatically publish safe metadata such as the record title, date, responsible agency, subject classification, document type, and a stable identifier. Sensitive metadata would be withheld or generalised according to clear rules.

• Make requests precise: applicants could identify a record instead of asking for “all documents relating to” a subject.
• Publish routine material automatically: final research reports, gift registers, organisational charts, ministerial diary data, and other low-risk categories could be released at creation.
• Prevent duplicate work: previously released documents would remain discoverable through a stable link instead of being processed again for each applicant.

This is valuable to journalists because it reveals the shape of government activity before a request is made. It is valuable to agencies because a precise record identifier sharply narrows the search task.

2. Give people secure access to their own records

Personal-information requests accounted for 32,109 of the 43,456 FOI requests received in 2024–25—74% of the total. These requests include immigration histories, welfare decisions, disability records, tax information, and other files that people need to navigate their relationship with government.

Many should not require a formal FOI case. Estonia’s “once-only” principle and X-Road data exchange demonstrate how government systems can share authoritative data securely while citizens access services through a digital identity. Australia can apply the same design principle through myGov without copying Estonia’s institutional model wholesale.

• Give authenticated users a clear inventory of personal records held by participating agencies.
• Allow immediate download of standard records that do not require exemption or third-party review.
• Show an access log so people can see which agency used their data, when it was accessed, and for what authorised purpose.
• Route complex or contested material into FOI with the request details already populated, preserving review rights.

Self-service should be an additional access channel, not a digital-only barrier. People who cannot use myGov, need an accessible format, or require support must retain equivalent assisted and offline options.

3. Make documents open by design

Redaction is expensive because sensitive and releasable information is often mixed together in unstructured files. Public sector templates should separate common information types at the point of authoring: personal details, legal advice, commercially sensitive figures, decision rationale, evidence, and approved public summaries.

Structured fields would let disclosure software prepare a proposed public version and explain which rule triggered each redaction. Machine-assisted review could also flag likely personal information in older files. A trained decision-maker would remain accountable for the release and for every exemption applied.

This distinction matters. The goal is not an opaque algorithm that decides what the public may see. The goal is to remove repetitive masking work, make redaction decisions auditable, and give human reviewers more time for difficult public-interest judgements.

Cost the opportunity without overstating it

The source policy model identifies large potential savings in the three most labour-intensive parts of FOI. Those scenarios are useful for setting pilot targets, but they overlap: a personal-record request removed from FOI also avoids search and redaction work. Their savings should therefore not be added together and presented as a budget forecast.

These figures are illustrative scenarios, not independent or additive savings forecasts. A business case should establish a common baseline, account for implementation and ongoing security costs, and publish sensitivity ranges.

What changes for the people who use the system

For journalists

A searchable register reveals which records exist and makes precise requests possible. Anonymous, no-cost searching protects reporters who are developing a story. Stable links to released records let readers inspect the evidence themselves.

For politicians and their staff

Proactive records create a shared factual base for debate and make it easier to answer constituents. Ministers also gain a more credible way to demonstrate that disclosure is routine and impartial, rather than selectively granted when politically convenient.

For public servants

Better metadata reduces broad manual searches. Personal self-service removes repetitive casework. Structured documents shorten review without removing professional judgement. Most importantly, clear release rules make openness an ordinary part of good recordkeeping instead of a last-minute legal emergency.

Build confidence through staged delivery

This reform should begin with services where volume is high, records are already digital, and disclosure rules are comparatively clear. A staged program would let government prove the value while testing privacy, security, accessibility, and operational impacts.

1. Set common rules: agree on a public metadata standard, protected metadata categories, stable identifiers, and minimum audit requirements.
2. Run bounded pilots: test one high-volume personal record service, one department’s correspondence register, and one structured policy-document workflow.
3. Publish independent results: report costs, processing time, accuracy, accessibility, incidents, user satisfaction, and human overrides.
4. Scale reusable infrastructure: expand only the components that demonstrate safer, faster, and cheaper access.

Privacy impact assessments, threat modelling, records-management review, and consultation with journalists, archivists, disability advocates, information officers, and affected communities should be part of design—not approvals sought after a platform is built.

Measure trust through observable service quality

Trust cannot be reduced to one survey score. Government can, however, publish service measures that show whether the system is becoming more dependable and open.

Make openness the cheaper default

A government that wants public trust should make its evidence easy to find, its decisions easier to inspect, and personal information easy for each person to access. That requires investment in public information architecture, but it also removes the recurring labour created by today’s fragmented systems.

The strategic choice is not between transparency and efficiency. Public record registers, personal data access, and open-by-design documents can improve both. By lowering the cost of release instead of raising the cost of asking, Australia can give journalists, politicians, public servants, and citizens a stronger reason to trust the machinery of government.

Sources and further reading

• Office of the Australian Information Commissioner: Annual report 2024–25
• Office of the Australian Information Commissioner: Charges for providing access
• Parliament of Australia: Freedom of Information Amendment Bill 2025
• Norwegian Government: eInnsyn in the Open Government Partnership action plan
• Norwegian Ministry of Digitalisation and Public Governance: 2026 evaluation of eInnsyn
• e-Estonia: X-Road and the once-only principle