AI safety

Dieses Rechtsdokument wird auf Englisch geführt.

Short version: in everything Boreon builds, a person holds the pen. Our AI reads, explains, and proposes. It does not act on your systems without a human approving the change, it is never trained on your data, and every product we ship works without AI at all.

Our position

We build AI the way we build networks and backups: as an instrument that serves people, never the other way around. That conviction is practical before it is philosophical. A tool that acts without approval is a tool your auditors cannot trust, and a tool trained on your data is a tool your lawyers cannot sign off. So we made the safe architecture the only architecture we ship.

What we build in

• A person approves every change. On the Boreon Platform, the AI drafts and a human approves. Changes route through the same approval gate your team already trusts, and the record of who approved what is kept.
• Read-only by construction. Our platform connectors carry no write method at all. A read-only credential is the most access they can ever hold, which is a property of the code, not a promise in a policy.
• Your keys, your model. You bring your own Claude, GPT, or OpenAI-compatible endpoint. We do not sit between you and your model provider, and we do not resell your tokens.
• No training on your data. Nothing you run through our products is used to train any model, ours or anyone's.
• AI is always optional. Every Boreon product has a deterministic path that works with no AI configured. If you unplug the model, the product still does its job.
• Confined where it counts. Our education tools keep AI inside the curriculum it was given, and our church tools keep a pastor, not a model, in the pulpit.

What we never do

• We never let AI execute changes to your environment on its own.
• We never train on customer data, and we never sell it.
• We never claim a certification we do not hold, or an accuracy we have not measured.
• We never make AI a requirement for a product to be useful.

Why we care

We are a Christian company, and we believe technology is only good when it serves human dignity. That is the standard we hold our AI to, and it is the standard we ask of the partners we build on.

Questions or concerns

If you believe an AI feature in any Boreon product has behaved outside these commitments, write to privacy@boreon.com. Reports about safety go to the front of the queue.