News Details
2026.02.27
Latest Trends in Regulations and Guidelines on the Use of Generative AI
1. Introduction
Generative AI—AI systems that automatically create text and images—has contributed significantly to business efficiency and the development of new services. At the same time, risks such as misinformation, copyright infringement, misuse of personal data, and discriminatory outputs have been widely discussed. As a result, countries around the world are rapidly developing regulations and guidelines governing the use of generative AI.
This article outlines the latest developments in major jurisdictions and highlights key practical considerations that companies should keep in mind.
2. EU: Phased Implementation of the AI Act
In the European Union, the AI Act entered into force on August 1, 2024. Its provisions are being applied in stages:
• Prohibited practices and AI literacy obligations: applicable from February 2, 2025
• Obligations related to General-Purpose AI (GPAI): applicable from August 2, 2025
Generative AI systems are likely to fall within the scope of GPAI. Companies may therefore be required to ensure transparency, establish risk management systems, and provide explanations regarding training data. Businesses connected to the EU market should review their contracts and internal rules accordingly.
3. Japan: Practical Governance Through Guidelines
In Japan, rather than introducing a comprehensive new law with uniform regulation, the government has adopted a guideline-based approach to clarify practical considerations. The “AI Guidelines for Business (Ver.1.1)” issued by the Ministry of Economy, Trade and Industry and the Ministry of Internal Affairs and Communications are designed to be continuously reviewed and updated.
From a practical perspective, companies should clearly document internal rules regarding:
1). The scope of information that may be input into AI systems;
2). Methods for verifying outputs;
3). Copyright and citation rules; and
4). Contractual arrangements with external vendors.
4. US and UK: Implementation Through Standards and Reporting Obligations
In the United States, discussions are ongoing at the federal level, while state-level initiatives are also advancing. In December 2025, New York State signed legislation requiring certain developers of so-called “frontier models” (highly advanced large-scale AI models) to disclose safety protocols and report major incidents within 72 hours.
In addition, the National Institute of Standards and Technology (NIST) has published the “AI RMF Generative AI Profile (AI 600-1),” providing a structured framework for implementing concrete risk management measures.
In the United Kingdom, rather than enacting a single comprehensive AI law, the government is promoting safe and responsible use through practical guidance such as the “AI Playbook.”
5. Conclusion
In this newsletter, we have reviewed the latest trends in regulations and guidelines concerning the use of generative AI. Regulatory approaches differ significantly across jurisdictions, including the EU, Japan, the United States, and the United Kingdom, both in philosophy and in implementation methods. As a result, companies face increasing difficulty in determining which jurisdiction’s standards should serve as the baseline for compliance and which issues may become significant in the future.
At ARK, we utilize insights obtained through the AGA network regarding regulatory developments and practical issues in various countries to provide structured perspectives and analytical frameworks to support companies in their ongoing assessments.
Please note that this article is intended to provide a general overview and does not constitute professional advice. We strongly recommend that you refrain from making decisions or taking action based solely on the information provided herein without consulting a qualified professional. Should you have any questions, please feel free to contact us.
Contact | Ark Outsourcing Co., Ltd. (ark-outsourcing.jp)
https://www.ark-outsourcing.com/jp/contact/
References
• EU “AI Act”
(Accessed February 26, 2026)
• Ministry of Economy, Trade and Industry, “AI Guidelines for Business”
(Accessed February 26, 2026)
• NIST, “AI RMF Generative AI Profile (NIST-AI-600-1)”
(Accessed February 26, 2026)
• UK Government, “AI Playbook”
(Accessed February 26, 2026)
Query on the phone+81-3-6821-9455