AI & innovation
Responsible AI for public and regulated services
If your service makes or informs decisions about people, AI brings obligations as well as opportunity. We help teams meet them without grinding delivery to a halt: data protection impact assessments, transparency, bias testing and the evidence an assessment panel or regulator will ask for.
We work to the UK Service Standard, the Algorithmic Transparency Recording Standard and the ICO’s guidance on AI and data protection, and we have taken services through GDS alpha and beta assessments.
What we cover
- Purpose and necessity. Is AI the right tool for this decision, and what is the human alternative?
- Data protection. DPIA support, lawful basis, minimisation and retention for training and inference data.
- Fairness. Testing outcomes across user groups, including people with accessibility needs and those in vulnerable circumstances.
- Transparency. Plain-language explanations for users, records for the transparency register and meaningful ways to challenge a decision.
- Operations. Monitoring, quality review and escalation once the service is live.
What you get
A responsible AI plan written for the service you are building, the artefacts you need for assessment, and a user-facing explanation that has been tested with real people.
Organisations we have done AI and innovation work with
Build AI into a public service responsibly
We can review a service you are planning or one that is already live.