From harm vocabulary to proposed Australian legal categories

Bounded taxonomy comparison, 9 September 2026, after the recorded forest assessment. This note compares express source use with legal choices; the September 2026 document remains an exposure draft. It does not establish that WEF, HRLC or Reset controls Australian drafting or enforcement.

The strongest new finding is an independent official route for the WEF vocabulary before HRLC's submission. The government's April 2024 review issues paper already cited the typology. HRLC therefore cannot be treated as the sole intermediary. Independent official route here means an official citation before that HRLC submission; it does not establish causal independence from all advocates, earlier relationships or shared workstreams. The strongest narrowing is the published review's decision to defer a broader consolidation of financial, societal and misinformation harms. The draft has a materially different structure from the typology. Neither conceptual overlap nor omission of a label, by itself, resolves the draft's complete reach.

Express transmission and its limits

Sources: [Government issues paper, Part 3/fn17 and later fn88/93](http[local research file] [HRLC155, §§4.3–5.1.5](http[local research file] [eSafety guidance, types of online harms](http[local research file]

The separate [producer packet]([local research file] establishes the workstream and contextual limits from its inspected sources. Its original WEF PDF was not acquired; detailed counterpart definitions must retain the Resolver/explainer attribution. Do not promote that packet into a claim that every cell of the original matrix was visually inspected. This lane adds actual government-side use to the production/NGO path; it does not establish who selected the later statutory clauses.

Reset's broader model and the review's narrower choice

Reset's April 2024 Green Paper describes its own proposals as a moment-specific contribution to discussion. Printed p6 explains its focus on online safety and misinformation/disinformation. The five-pillar table, printed p7/PDF9, distinguishes duty, assessment, mitigation, transparency and accountability. Its proposed assessments encompass existing Australian content/abuse categories, scams, electoral processes, public security, rights, public health and environmental risks. It expressly draws comparisons with the EU DSA, UK OSA and proposed Canadian legislation; its commentary criticises plural duties for weakening a systemic focus. This is a documented set of proposed policy choices with identified antecedents, not a WEF-only derivation. [Reset Green Paper, cover/about and printed pp6–7](http[local research file]

In the published updated statutory review, recommendation 5, printed p56, proposes broad families concerning young people, mental/physical wellbeing, harmful practices, national security/social cohesion and other illegality. Its adjoining discussion says broader financial, societal and misinformation harms fit a duty model but are handled by other existing or proposed Australian legislation and regulators. It leaves possible consolidation under an expanded commission to a future review because the current reform is already substantial. This is a deliberate scope choice. It is not a named rejection of WEF or a finding those harms are imaginary. Printed p12 also frames risk assessment through EU/UK models and existing Australian industry codes. [Published review, printed pp12 and 56](http[local research file]

The government's April 2026 response supports recommendation 4 and supports recommendation 5 in principle. Its explanation defines that state as support for intent/merit with alternative delivery or further implementation consideration; it is not acceptance of every proposed term. For recommendation 5 it stresses vulnerable people, criminal activity and public safety alongside expression, and says the duty complements the Criminal Code without changing obligations under existing discrimination statutes. It separately declines to simply lower adult-abuse thresholds, citing operational costs and the distinction between systemic prevention and reactive schemes. [Government response, PDF/printed pp3–6, recommendations 4–5 and adult-abuse discussion](http[local research file]

A consequential report-version qualification

FOI25-247 Document 1, MS24-002100, is a departmental brief seeking acceptance of an updated final report. It records that Rickard sought final improvements at a 1 November 2024 meeting, after delivery on 31 October. Assistant Secretary Andrew Irwin cleared the brief on 8 November. Attachment C, PDF7, reports five additions and 68 total recommendations, with four characterized as already in the text but not called out; it lists new recommendation 5. Its introductory fifth item describes a daily penalty for non-compliance with a duty action, while item g/new recommendation35 describes a $10m company maximum for removal-notice non-compliance and says this was not previously addressed. These descriptions are inconsistent. The later government response says the published review has 67 recommendations. No original total, deleted recommendation or final numbering is reconstructed by subtraction. PDF8 records a terminology-order change placing mental before physical wellbeing. The department describes changes as consistent with the delivered report and not changing findings; this is its characterization, not an acquired complete version comparison.

This establishes a post-delivery recommendation call-out and editorial revision. It does not establish that the substantive harm analysis originated in November, identify the individual drafter of recommendation 5, or show that the minister authored it. The readable approval alternatives/date fields are not proof of executed ministerial acceptance. Use published updated review, and distinguish original delivery from revision, rather than silently dating every public-page formulation to 31 October. [FOI25-247, Document 1 PDF1–3 and Attachment C PDF7–8](http[local research file]

Operative comparison: initial categories, future additions and adult choice

The exposure draft's distinctions are consequential:

These are proposed Australian legal mechanisms, not an incorporation of WEF's matrix. False information can intersect listed conduct; future additions remain possible. [Exposure draft, §§25B–26A and 27B, printed pp26–34/37, PDF30–38/41](http[local research file]

Earlier Australian legal sources and competing explanations

The as-made 2021 Act already defined serious physical/mental harm, distinguished serious distress from ordinary emotional reactions, and imposed targeted child/adult abuse tests. Its adult test combined likely intent to cause serious harm with an ordinary-reasonable-person assessment of the material. Sections 8–9 addressed contextual merit/consent and abhorrent violence. Section 106 tied class 1 to the Classification Act 1995 refusal category. These are directly inspected predecessors to the 2023 typology, rather than a presumed earlier neutral vocabulary. They supply a concrete competing source for similarities, while not proving every draft clause was copied from them. [As-made Online Safety Act 2021, §§5–9 and 106](http[local research file]

The April issues paper also discusses Australian research and existing abuse/content schemes beside its WEF references. This is consistent with synthesis from several inputs. The record supports influence through framing and documented advocacy; it does not support an exclusive linear chain in which a private taxonomy became law unchanged. Conversely, the existence of older law and contextual safeguards does not determine how broad future instruments or enforcement would become.

Candidate relations and remaining discriminating record

Preserve WEF typology → government April issues paper: cited for descriptive framing, independently of WEF typology → HRLC155: expressly adopted framing. Add WEF typology → eSafety guidance: explicitly used for educational categorisation, with the page's update-date boundary. Preserve Reset model → HRLC155: expressly endorsed, and review recommendation 5 → government response: supported in principle. Treat review taxonomy ↔ draft categories as a documented comparison unless a clause-level attribution is obtained. None of these is a donor-instruction, private veto or enacted-ban relation.

The next discriminating record is a category-by-category policy instruction, consultation disposition or drafting table for §§25C–G, showing source, alternatives, threshold selection and ministerial clearance. The likely policy custodian is the department's online-safety reform team; the relevant record family is the 2025–26 duty-of-care drafting/consultation file, including instructions to legislative drafters. It would distinguish accepted wording from shared vocabulary and establish why broader labels were omitted or particular additions delegated. No such instruction or disposition was acquired here. This lane stops at that record-to-decision breakpoint for the forest review.

Actual reading scope and access state

All current access observations are 9 September 2026. This note contains scoped public-reader/indexed-primary findings. No original PDF bytes, screenshots or original-byte hashes were acquired in this lane. PDF/printed numbering is differentiated where verified. The separate prior authority note supplies a fuller clause discussion, not new visual evidence.

No contributor biographies, new donor roster, private accounts, outreach, publication or canonical case edits were added. No inference about unseen draft material or access restrictions follows from these text-only readings.