On January 26, 2026, the Japan Association of New Economy submitted its comments on the Cabinet Office’s draft “Principles and Code on the Protection of Intellectual Property and Transparency for the Appropriate Use of Generative AI (provisional title)”

【Summary of Comments】
1. Comments on General Remarks
　• Clarification of Legal Status and Utilization of Existing Frameworks
　　The legal status of this Code should be clarified. Furthermore, instead of establishing a new, independent reporting format, the use of existing disclosure frameworks, such as Corporate Governance Reports, should be permitted as a substitute.
　• Implementation of “Comply or Explain”
　　Since the specific details and standards for the reasons for non-compliance (“explain”) remain unclear, the presentation of a collection of best practices for both “comply” and “explain” cases should be considered.
　• Optimization of the Scope of Application
　　Requiring “Generative AI Providers” who offer services using third-party models via APIs or other means to disclose details of training data that are impossible for them to grasp would impose an excessive burden and be ineffective; therefore, disclosure should be limited to the scope that they can internally control and ascertain.

2. Comments on Principles 1, 2 and 3
[Principle 1] Measures for Ensuring Transparency and Protecting Intellectual Property Rights
　• The disclosure of granular information, such as crawler identifiers and architectural details, would lead to the leakage of sensitive competitive information and hinder a fair competitive environment. Furthermore, due to concerns over potential security risks, these items should either be removed or explicitly permitted to be withheld from disclosure.
　• The granularity of disclosed information including the status of policy compliance, should be determined by the economically rational judgment of businesses exercising autonomous governance, and implemented in a manner that avoids imposing excessive burdens.
[Principle 2] Responses to Individual Disclosure Requests
　• To ensure that this Code does not become a mechanism for circumventing or substituting judicial procedures, disclosure requests should be left to existing legal procedures through the courts to prevent potential abuse.
　• It should be clearly stated in the text that when information leakage could lead to irreversible damage, such as when trade secrets are involved, refusal to respond is permitted from the perspective of protecting the legitimate interests of businesses.
[Principle 3] Responses to Inquiries for Specific Contents
　• The mere presentation of outputs and prompts is insufficient to determine the legitimacy of an inquirer, and since there is a risk of permitting exploratory requests (“fishing expeditions”) for the purpose of information gathering without sufficient legal grounds, objective substantiation should be required.
　• Information on whether specific URLs are included in training data constitutes the core of a business’s data collection know-how, and since a uniform obligation to respond poses a risk of leaking trade secrets through reverse engineering, the refusal to respond should be permitted at the business’s discretion.
　• Responding to individual disclosure requests requires significant time and resources and hinders technological innovation by startups with limited resources; therefore, mitigation measures such as phased implementation or the development of simplified response processes should be considered.

Please click here for the full text of the submitted comments. (Japanese only)