Responsible AI
We use AI to build faster and better — never recklessly. Here is how we keep your project, your data, and your rights safe when AI is in the loop.
Last updated: 2026
This page explains how Scriptus uses artificial intelligence in the work we do — for clients, and on our own products. It is written plainly; the smaller notes add the technical detail for those who want it.
A human owns every line
We use AI to work faster and explore more options — but a senior engineer reviews, understands, and is accountable for everything that ships. AI is a power tool in experienced hands, never a replacement for them. It does not commit to your project unsupervised.
Typed contracts, automated tests, and human code review on every change. No AI-generated code reaches your main branch without a person reading it, understanding it, and signing off.
Your code and data stay yours
We do not feed your code, your data, or your business into public AI tools that could train on it, leak it, or reuse it for someone else. What is yours stays yours.
We use tools and settings that exclude your content from model training, scope each tool to only what it needs, keep secrets out of prompts, and — where the work runs in your own AWS account — keep your data inside your own boundary.
Agents run on a leash
When we use AI agents to automate parts of the work, they operate inside firm guardrails. Anything that could matter gets a human checkpoint before it happens, and everything is recorded.
Least-privilege permissions, sandboxed execution, explicit human approval gates for risky or irreversible actions, and a full audit trail of what ran and why.
We harden the AI we build for you
If your product uses AI, we build it to resist misuse from day one — so a clever prompt cannot make it leak data, run up your bill, or behave in ways you never intended.
We design against the known LLM risks (the OWASP LLM Top 10): prompt-injection defenses, validation of what goes in and what comes out, rate and spend limits, content filtering, and least-privilege access for anything the model can trigger.
AI media, used properly
When we use AI to generate images, audio, or video for your project, we use properly licensed tools and keep you clear of rights problems. If you would rather have human-made media, we will arrange it — just ask.
We use commercial-tier tools that grant rights to the output, keep a record of how each asset was produced, and avoid tools with unclear or disputed training-data rights for anything we hand you as a deliverable.
We tell you where AI is used
No hidden automation and no overstated capabilities. We are straightforward about where and how AI is part of your project, and what it can and cannot do. If you would prefer we not use AI for some part of the work, that is your call to make.
Questions about how we use AI
If you have a question about our use of AI on your project — or you want a part of the work done without it — email us at hello@scriptus.ca. A real person reads it.