Public version log — Digital Tourism Think Tank
Versions and schemas
The complete public changelog for all four DTTT AI Transparency Framework models, the combined framework page and the AI Transparency Card toolkit. JSON schema files for each published version are available to download.
This page is the authoritative public record of how the framework has developed. It is updated at every published release. Version records note what changed, why, and what status each version holds. Organisations citing the framework in contracts, publications or procurement documents should reference the specific version number recorded here.
All JSON schemas are published under Creative Commons Attribution 4.0 International. You are free to use them in integrations, tooling and research with attribution.
Introduces a three-option entry flow with binary declarations at 0% (Human Created) and 100% (Fully AI) sitting outside the A to E scale. Adopts symmetric boundary structure: grades A and E each cover 10 scores at the extremes; grades B and D each cover 26; grade C covers 27. Grade A renamed from Human Created to Human Led. Grade E renamed from Fully AI to AI Led. Scope definition added. Grading consistency guidance and standard prompt template added.
- Entry flow: three-option selector replaces direct slider
- Declarations: Human Created (0%) and Fully AI (100%) introduced outside the A to E scale
- Grade A: renamed Human Led; range changed from 0–10% to 1–10%
- Grade B: range changed from 11–30% to 11–36%
- Grade C: range changed from 31–55% to 37–63%
- Grade D: range changed from 56–80% to 64–89%
- Grade E: renamed AI Led; range changed from 81–100% to 90–99%
- Scope definition added distinguishing generative AI tools from grammar and search tools
- Detailed Declaration Card format introduced for strategic deliverables
- Grading consistency section added with standard prompt template
Initial public release. Establishes the A to E grading scale with 0 to 100% banding, the Transparency Report Card structure, the self-assessment prompt and the public page. Superseded by v1.1.
Delivery Extension Score made external and member-facing. Previously internal only. Combined AI Value score added as a summary score derived from a 5x5 lookup matrix. Self-assessment tool rebuilt as a unified two-column interface showing all three scores simultaneously.
Replaced AI Augmentation Score with Delivery Extension Score following team and committee feedback. Augmentation framing positioned AI as the expert rather than the team. Delivery Extension Score was internal-only at this version. Superseded by v1.2.
- AI Augmentation Score removed from external Report Card entirely
- Delivery Extension Score introduced as internal-only replacement
- Five Delivery Extension levels redefined around delivery capability extension
- Report Card fields updated to remove Augmentation Score field
Initial release. Established Productivity Score (1 to 5, external) and AI Augmentation Score (1 to 5, external) as two independent scores on the Report Card. Superseded by v1.1.
Grade labels renamed to energy intensity framing throughout. Disclosure language standardised across all five grades to encourage consistent disclosure at every level, not only at higher grades.
- Grade A label: Negligible impact → Low energy intensity
- Grade B label: Low impact → Low-moderate energy intensity
- Grade C label: Moderate impact → Moderate energy intensity
- Grade D label: High impact → High energy intensity
- Grade E label: Very high impact → Very high energy intensity
- Grade definitions rewritten: consistent Disclose instruction across all five grades
- Colour key updated: Green — negligible → Green — low intensity
Introduces the A to E grading scale with green-to-red colour direction, three-axis scoring logic (task type, model category, usage intensity) and expanded methodology sources. Mitigation reframed as a conscious choices disclosure field with no grade adjustment effect. Grade boundaries to be tightened at v1.0 as EU AI Act first-party disclosure data becomes available from August 2026.
- A to E grading scale introduced with grade definitions, colour values and typical use examples
- Three-axis scoring: task type, model category and usage intensity
- Base grade grid introduced mapping task type and model category to starting grade
- Usage intensity modifier: light moves one grade down, heavy moves one grade up
- Mitigation removed as grade modifier; replaced with conscious choices disclosure field
- Methodology sources expanded to include Google Gemini disclosure, Epoch AI, Hugging Face, MIT Technology Review and EU AI Act Code of Practice
Initial pre-release. Establishes scope and three energy intensity categories (Low, Medium, High). A to E grading scale deferred to v0.5. Superseded by v0.5.
DC-NONE standard field and description updated from ‘unacceptable’ to ‘not_recommended’ to align consistently with the four-level Integrity classification scale.
- DC-NONE standard: unacceptable → not_recommended
- DC-NONE description: ‘Unacceptable for any visual or persona content’ → ‘Not recommended for any visual or persona content’
Initial release. Establishes a three-axis assessment structure covering Intervention (what AI did), Consent (what authorisation exists for real subjects) and Disclosure (what the audience was told). Produces a four-level Integrity classification (Clear, Caution, High Risk, Not Recommended) and a three-part machine-readable disclosure code.
- Six Intervention codes: AI-TXT, AI-IMG, AI-ENH, AI-ANIM, AI-CHAR, AI-PERS
- Six Consent codes: CS-NA, CS-ANON, CS-AI, CS-STD, CS-NONE, CS-EST
- Four Disclosure codes: DC-PROM, DC-CAP, DC-META, DC-NONE
- Classification logic defined for all four levels
- Disclosure code format defined: AI-[type] · CS-[consent] · DC-[disclosure]
- Procurement clause template provided
- Open governance question noted: whether AI-PERS can ever reach Clear classification
Environmental grade labels updated to energy intensity framing across all model tab references and the environmental self-grader. Disclosure language standardised across all five grades. Early adoption notice added to Overview tab with governance committee CTA. Eyebrow text updated to reflect four-model suite on all model sections. Report Card environmental label corrected.
- Environmental tab: grade labels updated to Low / Low-moderate / Moderate / High / Very high energy intensity
- Environmental self-grader JS: grade labels and definitions updated to match
- Overview tab: early adoption notice added with Design Week June 2026 reference and Typeform CTA
- Report Card: environmental badge label corrected from Moderate impact to Moderate intensity
- All model eyebrow text: updated from of 3 to of 4
Major editorial and structural rebuild. All model descriptions rewritten for industry-wide audience. Navigation restructured with Models and Toolkits dropdowns. New Research tab covering productivity measurement, academic research partnership programme and community participation. New How to apply tab with six implementation scenarios. Ethics Consideration Tool cross-referenced throughout. JSON schemas published. Transparency card on Report Card tab rebuilt to match toolkit Detailed Declaration Card output exactly.
Added Content Integrity Model as Model 4 with full tab including three-axis reference tables, classification scale, disclosure code block and live self-grader. Updated for four-model suite throughout. Superseded by v5.0.
Models at this build: Transparency v1.1, Productivity v1.2, Environmental v0.5, Integrity v0.1.
Updated Transparency tab to v1.1: three-option entry flow, Option B grade boundaries, Human Led and AI Led grade names, scope note and grading consistency guidance. Models: Transparency v1.1, Productivity v1.2, Environmental v0.5.
Added toolkit CTA block to Overview tab. Models: Transparency v1.0, Productivity v1.2, Environmental v0.5.
Productivity tab rebuilt: side-by-side score tables, unified self-assessment tool with Combined AI Value score. Models: Transparency v1.0, Productivity v1.2, Environmental v0.5.
Full tabbed redesign. Five tabs: Overview, Transparency, Productivity, Environmental, Report Card. Each model independently navigable. Models: Transparency v1.0, Productivity v1.1, Environmental v0.5.
Added Environmental Model v0.5 section. Single-page linear format. Models: Transparency v1.0, Productivity v1.1, Environmental v0.5.
Initial combined public page. Transparency Model and Productivity Model v1.0. Interactive slider included. Models: Transparency v1.0, Productivity v1.0.
Model selection opens with all four models unchecked — users choose which to score against before seeing any grading steps. Start Scoring button activates once at least one model is selected. Incomplete section validation added: sections missing answers display a clear warning before the card can generate. All internal links updated to full absolute URLs. Models: Transparency v1.1, Productivity v1.2, Environmental v0.5, Integrity v0.1.
Added Content Integrity as a fourth optional model. Integrated Integrity classification and disclosure code into all four embed formats. Scores strip rebuilt as CSS grid. Pre-release labels removed from Environmental and Integrity. Superseded by v3.0.
Three-option entry flow added to Transparency grader. Human Created declaration hides Productivity and Environmental sections. Detailed Declaration Card added as a fourth embed format with comprehensive AI tool and model checklists grouped by category and provider.
Initial release of the AI Transparency Card generator and embed tool. Models: Transparency v1.0, Productivity v1.2, Environmental v0.5.