AI Ethics

1. Our principles

These are the values that guide our choices when we build with AI. Where our Responsible AI page explains our practices, this page explains the principles behind them.

• AI should serve people and business outcomes, not replace human judgement.
• We are honest about what AI can and cannot do.
• We keep humans accountable for decisions.
• We prefer simple, useful automation over complexity for its own sake.

2. Fairness

We seek to avoid AI uses that produce unfair, discriminatory or harmful outcomes. We consider who is affected by an automation and design to reduce the risk of bias.

3. Accountability

A person — not a model — is accountable for outcomes. We keep clear ownership of AI-enabled systems and a way to trace, review and correct what they do.

4. Human oversight

Humans stay in the loop for meaningful decisions. AI assists; people review before outputs are relied upon.

5. Human accountability

This is the principle that anchors the others: AI assists, but humans remain responsible for business decisions. We do not present AI as a decision-maker. The people using AI-assisted outputs — our team and our customers — remain accountable for the choices they make with them. We design our services so that responsibility always rests with a person, not an algorithm.

6. Responsible deployment

We deploy AI only where it is appropriate and beneficial, avoid uses that could mislead or harm, and are willing to say no to a use case that conflicts with these principles.

7. How this relates to our other pages

Read this alongside AI Transparency (how AI is used) and Responsible AI (the practices we follow).