{ "model": { "id": "dttt-transparency", "name": "DTTT AI Transparency Model", "version": "1.1", "release_date": "2026-03-11", "release_type": "major", "status": "published", "description": "Classifies the balance of human and AI contribution to any deliverable. Uses a three-option entry flow: Human Created (declaration, 0%), AI Involvement (graded A–E, 1–99%), and Fully AI (declaration, 100%). The A–E grade scale uses symmetric narrow bands at the extremes (A: 1–10%, E: 90–99%) and an evenly distributed middle range.", "audience": "external", "url": "https://thinkdigital.travel/ai-transparency" }, "scope": { "definition": "AI involvement means the use of generative AI tools — tools that produce, draft, summarise, translate or transform content using a language or image model.", "in_scope": [ "Drafting or rewriting with ChatGPT, Claude, Gemini, Copilot or similar", "Research using an AI tool's knowledge or reasoning capability", "Image or video generation using AI", "Summarising documents with AI", "Translation via a generative AI tool", "Using an AI coding assistant to write or refactor code", "Prompt engineering and direction of AI output" ], "out_of_scope": [ "Spellcheck and autocorrect", "Grammarly and similar grammar correction tools", "Search engine results", "Predictive text on a keyboard or mobile device", "Stock image libraries with no generative element", "Template-based tools with no generative AI component" ], "note": "Prompt writing, selecting and supplying source documents, and directing AI output are all forms of human input and should reduce the AI involvement percentage accordingly." }, "entry_flow": { "description": "Users begin by selecting one of three options before reaching the grader.", "options": [ { "id": "human_created", "label": "Human Created", "type": "declaration", "percentage": 0, "description": "No AI involvement at any stage. Declared directly — no letter grade assigned.", "report_card_label": "Human Created (declared)" }, { "id": "ai_involvement", "label": "AI Involvement", "type": "graded", "description": "AI played some part. User proceeds to the A–E grader to set a percentage and receive a letter grade.", "range_min": 1, "range_max": 99 }, { "id": "fully_ai", "label": "Fully AI", "type": "declaration", "percentage": 100, "description": "No human authorship beyond the initial brief. Declared directly — no letter grade assigned.", "report_card_label": "Fully AI (declared)" } ] }, "grades": [ { "grade": "A", "label": "Human Led", "range_min": 1, "range_max": 10, "options_count": 10, "colour_hex": "#cde9e7", "text_colour": "#0b0c0c", "definition": "AI influence is negligible. AI may have been used briefly as a sounding board, for a single query, or for minor phrasing suggestions, but the work is substantially human in origin.", "typical_use": "In-person workshops or sessions where AI was consulted briefly before or after; reports where a single AI query informed background research.", "boundary_rationale": "Narrow by design. Human Led is a strong claim and should require genuine confidence. Only work with negligible AI use belongs here." }, { "grade": "B", "label": "AI-Assisted", "range_min": 11, "range_max": 36, "options_count": 26, "colour_hex": "#66bfb9", "text_colour": "#0b0c0c", "definition": "AI used for research, ideation or drafting support. The human retains full creative and strategic direction throughout.", "typical_use": "Keynotes where AI supports background research; strategic papers where AI assists with literature review or initial structure.", "boundary_rationale": "Wide enough to capture the range of ways AI can assist without co-authoring. Matched to D in width to keep the middle three bands symmetrical." }, { "grade": "C", "label": "Collaborative", "range_min": 37, "range_max": 63, "options_count": 27, "colour_hex": "#00857C", "text_colour": "#ffffff", "definition": "Balanced human and AI contribution. AI handles research, initial drafting and structural suggestions while the human directs strategy, edits and refines.", "typical_use": "Standard advisory reports, member briefings, strategy documents. The most common grade for DTTT and DMO advisory work.", "boundary_rationale": "The widest band, with one additional point, because this is where most advisory and knowledge work sits. More granularity in the middle is more useful. Grade C receives the extra point because 101 scores cannot divide into five perfectly equal bands." }, { "grade": "D", "label": "AI-Generated", "range_min": 64, "range_max": 89, "options_count": 26, "colour_hex": "#005c57", "text_colour": "#ffffff", "definition": "AI-led production with human review, quality control and final judgement. The human shapes direction but AI produces the bulk of the content.", "typical_use": "Data-heavy research compilations, rapid benchmarking requests, large-scale case study catalogues.", "boundary_rationale": "Matched to B in width to keep the middle three bands symmetrical." }, { "grade": "E", "label": "AI Led", "range_min": 90, "range_max": 99, "options_count": 10, "colour_hex": "#003835", "text_colour": "#ffffff", "definition": "Minimal human involvement. Output is substantially unedited AI generation; human contribution is limited to brief, direction and light review.", "typical_use": "Automated template outputs; AI-generated content reviewed without significant editing or direction.", "boundary_rationale": "Narrow by design, matched to A. AI Led is as strong a claim as Human Led. Only work where human involvement is genuinely limited to commissioning the brief should reach this grade." } ], "boundary_summary": { "total_scores": 101, "note": "101 is a prime number and cannot be divided into five perfectly equal bands. The symmetric approach assigns 11 scores to each declared extreme anchor (0% and 100%), 10 scores to each of grades A and E (1–10% and 90–99%), and distributes the remaining 79 as evenly as possible across B, C and D: 26, 27 and 26 respectively. Grade C receives the extra point.", "distribution": { "Human Created declaration": 1, "A Human Led": 10, "B AI-Assisted": 26, "C Collaborative": 27, "D AI-Generated": 26, "E AI Led": 10, "Fully AI declaration": 1 } }, "design": { "gradient": "linear-gradient(to right,#cde9e7 0%,#66bfb9 25%,#00857C 50%,#005c57 75%,#003835 100%)", "gradient_label_start": "Human", "gradient_label_end": "AI", "brand_colour": "#00857C", "declaration_human_bg": "#f7fefe", "declaration_human_border": "#cde9e7", "declaration_ai_bg": "#0d1f1e", "declaration_ai_border": "#003835" }, "report_card": { "formats": [ { "id": "standard", "label": "Standard Report Card", "description": "The default format for all deliverables. Includes grade, models used, tasks and summary fields.", "fields": [ "ai_involvement_grade", "models_used", "ai_supported_tasks", "ai_contribution_summary", "human_contribution_summary", "transparency_note" ] }, { "id": "detailed", "label": "Detailed Declaration Card", "description": "A comprehensive format best suited for significant strategic or public-facing deliverables such as destination strategies, new websites, major research reports and industry publications.", "fields": [ "deliverable_title", "deliverable_type", "date", "organisation", "ai_involvement_grade", "models_used", "ai_supported_tasks", "prompt_quality_note", "source_documents_provided", "ai_contribution_summary", "human_contribution_summary", "transparency_note", "value_breakdown", "productivity_score", "delivery_extension_score", "environmental_grade" ], "note": "The optional scores for Productivity, Delivery Extension and Environmental Impact are drawn from the other two models in the DTTT framework. Include them if the full framework has been applied." } ], "task_list": [ "Research", "Drafting", "Data analysis", "Brainstorming", "Structure", "Editing", "Visual mockups", "Prompt engineering", "Translation", "Fact-checking", "Code generation", "Presentation design" ] }, "grading_consistency": { "description": "Guidance for ensuring consistent grades whether the work is self-graded by a human or assessed by an AI model reviewing the deliverable.", "risks": [ "AI models may weight formatting or document structure heavily when assessing contribution, inflating the grade", "Grading a conversation or draft rather than the final deliverable produces an inaccurate result", "Without the scoring boundaries explicitly in the prompt, AI may apply different band definitions", "The Combined Value score matrix must be included in the prompt for the Productivity Model or AI will use inconsistent weightings" ], "standard_grading_prompt": "You are assessing AI involvement in a completed deliverable using the DTTT AI Transparency Model v1.1. Apply the following rules exactly.\n\nScope: AI involvement means generative AI tools only (ChatGPT, Claude, Gemini, Copilot, Midjourney and similar). Spellcheck, grammar tools, search engines and predictive text do not count.\n\nGrade boundaries:\n- Grade A — Human Led: 1–10%. AI influence negligible.\n- Grade B — AI-Assisted: 11–36%. AI used for research or drafting support; human retains full direction.\n- Grade C — Collaborative: 37–63%. Balanced contribution; both human and AI shaped the output.\n- Grade D — AI-Generated: 64–89%. AI produced the bulk of the content; human reviewed and directed.\n- Grade E — AI Led: 90–99%. Minimal human involvement beyond brief and light review.\n\nIf no AI was involved at any stage, record this as: Human Created (declaration, 0%).\nIf the work involved no meaningful human authorship beyond the brief, record as: Fully AI (declaration, 100%).\n\nAssess only the final deliverable provided. Do not assess formatting or document structure unless AI was explicitly used to design it. Focus on: content authorship, research origin, strategic direction, structural decisions and quality of human editing.\n\nProduce:\n1. AI Involvement Grade with percentage and one-sentence justification\n2. Models used\n3. AI-Supported Tasks (from: Research, Drafting, Data analysis, Brainstorming, Structure, Editing, Visual mockups, Prompt engineering, Translation, Fact-checking, Code generation, Presentation design)\n4. AI Contribution Summary (2–3 sentences)\n5. Human Contribution Summary (2–3 sentences — include prompt quality, source documents provided and editorial decisions)\n6. Transparency Note (one sentence for the deliverable footer)\n\nBe accurate. If AI contribution was limited, record it as limited.", "human_review_note": "Any AI-generated grade assessment should be reviewed by the person who produced the work before publication. The human author is best placed to confirm whether the grade accurately reflects how the work was produced." }, "governance": { "licence": "Creative Commons Attribution 4.0 International", "contact": "info@thinkdigital.travel", "committee": true, "versioning_policy": "Minor releases: refinements to language, definitions or examples. Major releases: structural changes to grade definitions, entry flow or scope. All major releases require committee review." }, "changelog": [ { "version": "1.1", "date": "2026-03-11", "type": "major", "summary": "Introduces a three-option entry flow with binary declarations at 0% (Human Created) and 100% (Fully AI) sitting outside the A–E scale. Adopts symmetric boundary structure (Option B): 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 to distinguish generative AI tools from standard productivity tools. Grading consistency guidance and standard prompt template added.", "changes": [ "Entry flow: three-option selector (Human Created / AI Involvement / Fully AI) replaces direct slider", "Declarations: Human Created (0%) and Fully AI (100%) introduced as declaration types outside the A–E grading scale", "Grade A: renamed from Human Created to 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 from Fully AI to AI Led; range changed from 81–100% to 90–99%", "Scope definition added: generative AI tools in scope; spellcheck, grammar tools, search and predictive text out of scope", "Boundary rationale documented per grade", "Detailed Declaration Card format introduced for strategic and significant deliverables", "Grading consistency section added with standard prompt template", "Colour values corrected in JSON to match published page: monochromatic teal scale" ] }, { "version": "1.0", "date": "2026-03-01", "type": "major", "summary": "Initial public release of the AI Transparency Model.", "changes": [] } ] }