AI Usage Policy
SHANNON LIMITED LIABILITY COMPANY (hereinafter, "we" or "the Company") establishes the following policy on the use of AI technologies, including generative AI, in our work. We recognize that while the use of AI technology improves the productivity of our work and the value we deliver to customers, it also requires care with respect to information security, copyright, and privacy. This policy sets out, in writing, the criteria we apply when using AI in our work, as well as the premises we adopt when proposing and implementing AI in a customer's environment.
1. Our Basic Stance on Using AI
We actively use generative AI across the design, implementation, and operation phases of our work. Specifically, we incorporate AI services on business plans whose input data is not reused for training—such as Claude by Anthropic—into our work and provide an environment in which our members can use them on a daily basis. By treating AI not merely as a supporting tool but as a premise of our work, we aim to achieve both quality and speed in the value we deliver. At the same time, our stance that a human always makes the judgment and bears the responsibility for AI output remains unchanged.
2. Scope
This policy applies to generative AI (large language models, image-generation AI, code-generation AI, and the like), machine learning services, and software that incorporates AI features in general, as used in our work. It also covers each stage of proposal, design, implementation, and maintenance and operation for customers, as well as our internal work (creating materials, code review, preparing documentation, and so on).
3. Principles of Use
We use AI based on the following principles.
- Handling of confidential information — We do not input confidential information, personal information, or non-public materials provided by customers into anything other than the business plans we have approved. Input into services where data may be reused as training data is prohibited.
- Human review — Deliverables generated by AI are always reviewed and verified by a human before being used in our work. For final deliverables provided to customers, we deliver what our members have reviewed and revised.
- Ensuring transparency — When AI output is used in a final deliverable for a customer, we make this fact and the verification process clear as necessary. When a customer asks us to explain our usage policy, we disclose it candidly.
- Confirmation of rights — For output derived from the copyrighted works or trademarks of third parties, we use it only after confirming the rights involved. If we determine that the output may infringe the rights of others, we do not use it.
- Incident response — If there is a risk of a security incident or information leakage, we promptly suspend use, investigate the scope of impact, and report to the customer as necessary.
4. Use of AI in Customer Environments
When proposing and implementing AI services in a customer's environment, we follow, in addition to this policy, the customer's compliance requirements, contractual terms, and industry regulations. Specifically, we decide the following on a case-by-case basis through individual consultation.
- The type and contractual form of the AI services used (whether business plans that are not reused for training are required or recommended, and to what extent)
- The scope of data that may be input and how it is handled
- The verification process for AI output and the allocation of responsibility
- The retention policy for audit logs and usage history
5. Responsibility and Verification
Ultimate responsibility for AI output rests with our members and with the Company. We take responsibility for verifying AI-generated deliverables before they are provided. We do not reuse information provided by a customer as training data without the customer's permission.
6. Revisions
We will revise this policy from time to time in response to changes in AI technology and related laws and regulations. Any revised version takes effect when it is posted on this page. In the event of a significant change, we may notify customers individually.