Our principles
Eight commitments that guide everything we publish, recommend and stand for.
- 01
People first.
Every AI system answers to the people it affects, not the other way round. Human dignity is the measure of whether a system is good.
- 02
Accountability is human.
Responsibility for what an AI system does always rests with a person or an institution. Software cannot be held to account; its owners can.
- 03
Transparency by default.
People have a right to know when they are dealing with AI, and how it shapes the outcomes that concern them.
- 04
Privacy is a right.
Personal data is an extension of the person it describes. It is not raw material to be collected and reused without limit.
- 05
Fairness is designed, not assumed.
Bias is measured, disclosed and corrected. A system that treats people unequally is not neutral because it is automated.
- 06
Freedom of thought.
No system should manipulate belief or behaviour at scale. Persuasion that people cannot see is not persuasion; it is control.
- 07
A right to a human.
Anyone affected by an automated decision can reach a person, understand the reasons and contest the outcome.
- 08
Coexistence, not replacement.
Intelligent systems should extend what people can do, share their benefits widely, and leave final control with humans.
Principles mean little until they are practised.
Help us turn them into positions, guidance and better systems.