This policy explains how Ryu Labs uses AI in our own work and in what we build for clients. It's the practical companion to our Safety page — this is the policy version, that's the plain-language version.
How we use AI internally
We use AI tools to assist with engineering, research, and drafting across our own workflow. Every AI-assisted deliverable goes through the same human review as any other work before it ships — AI use doesn't change our quality bar, it's a tool for reaching it faster.
AI in client deliverables
Where a client engagement includes AI features, we build and document an evaluation process specific to that feature before launch, and disclose to the client which parts of the product involve AI-generated or AI-assisted output.
Human oversight
We don't grant AI systems unsupervised authority over actions that are costly to reverse — financial transactions, irreversible deletions, or communications sent on a user's behalf — without the client explicitly requesting and accepting that trade-off.
Data & model training
We do not use client data to train or fine-tune foundation models for use outside that client's own systems. Where a project involves fine-tuning a model on client data, that model and its outputs remain scoped to that client's engagement.
Client disclosure obligations
If we build an AI feature for you that interacts directly with your end users, we'll recommend appropriate disclosure to those users as part of the delivery — the final decision on end-user disclosure sits with you as the product owner.
Limitations
AI systems, including ones we build, can produce incorrect or unexpected output. We design for that reality with evaluation, grounding, and human review rather than assuming it away, but we don't claim any AI system we build is error-free.
Contact
Questions about how AI is used in a specific engagement can be sent to ai@ryulabs.io.