AI content disclosure

Why this disclosure exists

Some editorial content on this directory is drafted with AI assistance. Where AI assistance has been used, the page carries a visible AI-content-disclosure block that identifies the assisted nature of the drafting and the named human reviewer who has substantively reviewed and approved the content.

The disclosure exists for three reasons. First, transparency: readers should know when AI has assisted in drafting so they can apply appropriate scrutiny. Second, trust: AI-assisted content can be high-quality but only when the underlying review is rigorous; the disclosure binds the reviewer's name to the published output. Third, search-engine credibility: external search engines (Google's Helpful Content Update guidance and equivalent from Bing, DuckDuckGo, etc.) consider AI-disclosed content with named-reviewer accountability favourably relative to AI-generated content without disclosure.

What "AI assistance" means here

AI assistance for editorial content typically takes one or more of the following forms: structuring an outline based on the page's hard-scope requirements; drafting initial paragraphs that the editor then revises; summarising authoritative sources for the editor to verify against the originals; cross-checking factual claims against the published source registry.

What it does not include: generating numerical claims that aren't independently verified; fabricating citations; producing text that the named reviewer has not substantively reviewed; replacing editorial judgement on what to cover and how.

Reviewer accountability

Every page on the directory has either: (a) a named human reviewer with their professional credentials listed in the page footer, who has substantively reviewed and approved the content, or (b) the byline_pending: true flag indicating the page is awaiting named-editor restamp and is currently noindexed pending that pass.

Pages in state (b) are visible to URL-direct visitors but carry the "Pending editorial review" banner and are excluded from search results. The directory's editorial cadence works through the byline-pending queue; the queue's status is visible in the methodology-overview article.

What the disclosure says

A page with AI-assisted drafting carries a disclosure block at the top of the article body with text similar to: "This article was drafted with AI assistance and reviewed by [named reviewer with credentials] on [date]. Source citations are linked inline; the editor has verified each numerical claim against the cited source."

The disclosure block is not collapsible; it is part of the article's standard layout and is visible without interaction.

What we do not use AI for

Review aggregation and moderation decisions involve human editorial review for any non-trivial outcome (review hides, reviewer-account flags, repeat-violation determinations). Automated moderation surfaces candidate decisions; humans confirm.

Firm-profile factual data (firm name, jurisdiction, partner count, credentialing-authority registrations) is sourced from public registries and firm-provided submissions, not generated by AI. Where ingest creates a firm-profile entry, the entry's data points are traceable to public-source records.

Sanctions-flag and credentialing-status data is sourced from authoritative registries. These data points are not generated by AI; they are pulled from registry APIs or scraped where APIs are not available, with human review of the integration logic.

Citations and source verification

AI-assisted drafts often need source verification: a claim that "country X has Y treaties" needs to be verified against a current authoritative source rather than against the AI's training data, which may be stale or incomplete. The directory's editorial workflow includes a citation-verification step where every numerical claim and named-source reference is independently verified by the reviewer against the underlying source.

Where a claim cannot be verified to a current authoritative source, the claim is removed from the article rather than published with the underlying uncertainty.

Impact on AEO and AI-search exposure

AI-assisted content with named-reviewer accountability is structurally different from AI-generated content without review, and external AI search engines (the major LLM-backed search products) are increasingly differentiating between the two via authority signals (named author with verifiable credentials, structured citations, freshness signals).

The directory's content strategy is built for the AI-search-enabled discovery layer: question-format H2s, 40-60-word answer paragraphs, FAQ blocks, structured-data markup, sourced citations, and named-reviewer accountability are all structural choices that improve content's exposure in AI-search responses while preserving the editorial standards that make the underlying content useful.

Reader feedback on AI-assisted content

If you encounter content that you believe is incorrect, misleading, or insufficiently reviewed, use the corrections-and-edits-policy submission flow. AI-assisted content is held to the same correction standards as other editorial content; corrections that succeed are applied and logged in the public corrections changelog.