Responsible AI – How We Think About It

If you haven’t read about our overall approach to AI yet, that’s a good place to start: Our AI Philosophy

This page lays out the specific commitments we’ve made about how AI works in our platform, the guardrails, the human checkpoints, and the boundaries we won’t cross.

Oil and gas operations carry real consequences. A drilling decision made on bad data doesn’t just cost money, it can cost time, equipment, and in the worst cases, safety.

We take that seriously in everything we build, and AI is no exception.

You generate it. You decide what to do with it.

When you ask for an AI-generated summary, it appears in a separate panel, not inserted into your report. You read it, decide if it’s right, and choose to use it, edit it, or ignore it entirely. Nothing goes into your data without you putting it there.

Deterministic decisions stay deterministic.

Anything in the platform that affects data integrity: calculations, allocations, record modifications. These will stay rule-based and deterministic. AI does not touch processes where a non-deterministic output could corrupt your data or produce results that can’t be traced. Where the line between advisory and operational isn’t clear, we build in human checkpoints until it is.

We’re building toward full auditability.

We believe every AI output should come with a visible trail…what data it drew on, what reasoning it followed, what it chose to include or leave out. We’re still designing exactly what that looks like in the interface, but the commitment is firm: no feature ships without a meaningful way to see how it reached its conclusion. Think less “black box” and more “showing its work.”

Your data doesn’t train anyone’s model, and that’s just the start.

Our API agreements explicitly prohibit training on client data. What happens in your platform stays in your platform. We’ve also built tiers of deeper privacy options for organizations that need more control. (More on our full privacy approach soon.)

We’re honest about what’s still being figured out.

We haven’t solved every ethics question in oil and gas AI. Data classification, edge case handling, the right level of AI involvement in different workflows…these are live conversations at Mi4, not closed ones. We’d rather tell you that than pretend we have all the answers.

You move at your pace.

Opt-in, always. No surprises. Every AI feature requires you to opt-in by default. Your administrator chooses if and when to turn it on for your organization, and even then individual users decide whether to use it. Nothing changes in your workflow until you actively want it to.

 

These commitments reflect where we are today. Some of them will evolve as the technology matures, as we learn from clients using these features in the field, and as we work through the questions we haven’t fully answered yet. We’ll update this page when our thinking changes…and we’ll tell you when we do.

If something here raises a question about your specific situation, we’d rather you ask it directly than wonder.

— Talal Nehme
Founder, Mi4 Corporation