Purpose before capability
We start with a defined need and a legitimate outcome, not with a tool looking for a use.
Responsible technology
Technology should extend human capability without obscuring responsibility.
We apply proportionate safeguards to the design, use and operation of automation, data-driven systems and AI.
Human accountability
Material decisions retain a responsible human owner. Automation has defined authority, escalation and override paths.
Privacy and data restraint
We minimise collection, access and retention, and consider privacy throughout system design.
Explainability and traceability
People should be able to understand when automation is involved, what it influences and how important outcomes can be reviewed.
Fairness and accessibility
We examine who may be excluded or disadvantaged and test with a suitable range of users and conditions.
Ongoing oversight
Performance, incidents and unintended effects are monitored after launch, with clear thresholds for intervention or withdrawal.
Practical governance
Principles become useful through controls.
Depending on context, controls may include data mapping, access reviews, human approval points, evaluation criteria, audit records, incident processes and periodic reassessment.
We do not present automated outputs as certain where they are probabilistic, and we do not remove human judgement where the consequences call for it.