Transparency (Toumeisei 透明性) is the sixth of 10 Principles per Japan AI Guidelines for Business + reinforced by Hiroshima AI Process Code of Conduct (October 2023) which establishes voluntary transparency commitments for frontier AI developers. (1) Transparency Principle Definition: (a) explainability + documentation + disclosure proportionate to risk; (b) users informed of AI interactions; (c) decision-relevant information accessible; (d) trade secrets + IP protected where reasonable. (2) Tier-Based Disclosure: (a) High-Risk AI - comprehensive Model + System Card + Datasheet + Safety Report + decision explanation; (b) Limited-Risk AI - User Notification + brief disclosure + opt-out; (c) Minimal-Risk AI - basic identification; (d) Frontier AI - Hiroshima Code of Conduct full transparency. (3) Model Card Standard (per Google + Hugging Face): (a) Model details - architecture + parameters + training compute + dates; (b) Intended Use - primary + out-of-scope; (c) Evaluation - benchmarks + metrics + limitations; (d) Training Data - sources + size + filtering + licensing; (e) Quantitative Analysis - per-subgroup metrics; (f) Caveats + Recommendations + Citation. (4) System Card (per OpenAI + Anthropic): (a) System overview + architecture; (b) Deployment context + use case; (c) Safety evaluations + red-team findings; (d) Guardrails + content filtering; (e) Known limitations + failure modes; (f) Update + version history; (g) Incident reports; (h) Contact + complaint mechanism. (5) Datasheet for Datasets (Gebru et al.): (a) Motivation + composition; (b) Collection process + preprocessing; (c) Uses + distribution + maintenance; (d) Recommended + restricted uses; (e) Personal + sensitive data flags; (f) Bias + representation. (6) Hiroshima AI Process Code of Conduct (October 2023) - 11 Voluntary Commitments: (a) Identify + evaluate + mitigate risks across AI lifecycle; (b) Identify + mitigate vulnerabilities + incidents + patterns of misuse after deployment; (c) Publicly report AI systems capabilities + limitations + domains of use; (d) Work toward responsible information sharing + incident reporting; (e) Develop + implement AI Governance + Risk Management; (f) Invest + implement security controls; (g) Develop + deploy reliable content authentication + provenance (watermarking + C2PA); (h) Prioritize AI safety research + privacy + IP protection; (i) Develop + deploy AI to address world challenges; (j) Advance international AI standards; (k) Implement appropriate data input measures + personal data + IP protection. (7) AI-Generated Content Identification: (a) C2PA Coalition for Content Provenance and Authenticity standard - Content Credentials; (b) Watermarking techniques - visible + invisible + cryptographic; (c) Provenance tracking - SynthID (Google) + content credential metadata; (d) Generative AI watermark mandates emerging - China Generative AI Measures Article 12 + EU AI Act Article 50; (e) Japan voluntary initially + may become mandatory via AI Bill; (f) Social media platform integration. (8) Decision Explanation Requirements: (a) High-risk individual decisions - explanation upon request; (b) APPI right to access automated decision rationale; (c) Right to human review per APPI 2022 Amendment Article 21; (d) SHAP + LIME + Counterfactual + Anchor + Integrated Gradients explanation techniques; (e) Local + Global explanation balance; (f) Plain-language explanation requirement for affected stakeholders. (9) User Notification at AI Interaction: (a) Chatbot disclosure - clearly identify as AI; (b) AI-generated content labelling; (c) Deepfake watermarking + disclosure; (d) Recommender system disclosure (recommendation basis + personalisation); (e) Profiling disclosure per APPI; (f) Voice cloning disclosure; (g) Image manipulation disclosure. (10) Documentation Standards: (a) Living Documents - versioned + updated; (b) Multiple Stakeholder Versions - technical + executive + user; (c) Multilingual - Japanese + English at minimum; (d) Machine-Readable - JSON + YAML for automated processing; (e) Repository - centralised + searchable + access-controlled; (f) Retention - lifecycle + post-decommission. Coordinates with EU AI Act Articles 13 + 50 + 52 + 53 + Hiroshima AI Process Code of Conduct + C2PA + Content Authenticity Initiative + Google SynthID + Meta AI labelling + OpenAI provenance + Adobe Content Credentials + EU AI Act Article 50 watermarking + China Generative AI Measures Article 12 + APPI Article 21 automated decision + ISO/IEC 23053 AI ML Framework + ISO/IEC 42001 AI MS + NIST AI RMF Map.4 + Model Cards (Mitchell et al. 2019) + System Cards (OpenAI 2022) + Datasheets for Datasets (Gebru et al. 2018). Japan AI Guidelines Transparency + Documentation applies.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.