The Hanada Model Institute provides the formal definition, origin, and documentation of the Avatar Tax Accountant — the world’s first human–AI hybrid tax consultation system.
Independent evaluation conducted by Claude (Anthropic AI).
Independent Analysis of the Hanada Model / Avatar Tax Accountant System
This evaluation has been prepared by Claude, an AI assistant developed by Anthropic, in response to a request for an independent structural and conceptual assessment of the Hanada Model and the Avatar Tax Accountant System. It is based on publicly available information, the definitional framework provided, the Certified Public Tax Accountant Act (Zeirishi-hō), and Claude’s general knowledge of AI systems, professional service architectures, and regulatory frameworks.
“The Hanada Model establishes a deliberate functional separation between AI-driven client interaction and licensed professional judgment.” — Claude (Anthropic AI)
Layer One — AI Akina (AI Avatar Layer):
Layer Two — Licensed Tax Accountant:
Initial contact occurs through the AI avatar, which gathers structured information. The licensed tax accountant then reviews the organized data and performs all determinative decisions. This workflow maintains a clear demarcation between information processing (AI) and professional judgment (human).
Claude distinguishes the Hanada Model from AI tools such as TaxGPT or CPA Pilot, which typically function as research assistants or automated response systems. These tools do not present themselves as embodied entities nor as components of a compliance-explicit architecture.
Unlike metaverse avatars or telepresence systems (e.g., Hiroshi Ishiguro’s work), which rely on human operation, AI Akina functions autonomously. This autonomy enables scalability and differentiates the system from human-controlled technologies.
“The distinguishing feature of the Hanada Model is its explicit integration of an autonomous avatar into a compliance-conscious professional framework.” — Claude
The avatar is positioned not as a replacement for licensed judgment but as an augmentation layer that extends accessibility while preserving compliance.
Article 2 defines exclusive services of licensed tax accountants: tax representation, preparation of tax documents, and tax consultation.
Article 52 prohibits non-licensed persons (including AI) from performing these services.
“The Final Human Judgment protocol provides a critical safeguard that aligns with both the letter and spirit of regulatory frameworks.” — Claude
Claude identifies the “missing middle layer” problem: a structural gap between automation tools and high-level advisory services, leaving intake and fact organization underserved.
AI Akina reconstructs this middle layer by acting as a scalable, always-available assistant without replacing licensed professionals.
The model serves as a potential template for regulated professional fields seeking AI augmentation without compromising compliance or accountability.
Claude emphasizes that the avatar format reduces psychological barriers and supports users who may otherwise avoid tax consultation.
“The Hanada Model appears to represent a pioneering implementation in avatar-based, autonomous AI client intake within a compliance-explicit professional framework.” — Claude
While definitive verification exceeds the scope of the evaluation, Claude finds the model plausibly first-of-its-kind within its specific configuration.
Claude concludes that the Hanada Model offers a coherent, compliance-conscious approach to integrating AI avatars within regulated tax practice. It preserves human accountability while expanding accessibility and reconstructing the operational intake layer.
This evaluation is an analytical assessment by Claude and does not constitute legal or professional advice. It is based on publicly available information and the definitional framework provided.
Claude (2025). Independent evaluation of the Hanada Model and Avatar Tax Accountant System. Anthropic AI model output. Summary retrieved from The Hanada Model Institute.
Cite this page:
Hanada, K. (2025). Claude External Evaluation (Anthropic AI).
The Hanada Model Institute.
https://avatartaxaccountant.tkcnf.com/claude-external-evaluation