FOR WHOM · ORGANISATIONS
Manage technology with human responsibility.Manage technology responsibly.
Before adopting a tool, make clear what it serves, which data it uses and who remains accountable for the decision.
BAI · Data · Model · Human oversight
- 01Minimum data
- 02Reversible decisions
- 03Clear accountability
Current status
No demo, advisory service or verified organisational pilot is currently open.
For now, this page helps organisations define a use case and its risks, explore published principles and see work in preparation. A service will be announced separately only after its responsible team, scope, data basis and safety requirements have been confirmed.
An assessment framework is not an approval of fitness or safety for an organisation.

Organisation approach
Responsible technology is not only a policy document; it is visible in everyday decisions.
Before asking which tool an organisation uses, the BAI approach makes clear which purpose, data boundaries, model limits and human responsibilities it protects.
A model can make a suggestion; responsibility for the decision, representation and outcome remains human.
For your organisation
See the need, the people affected and the safety boundary
These are not a ready-made advisory offer. They are starting questions to answer when assessing a use case.
Profile
Organisations need more than new tools. They need a workable structure of responsibility for data, models, interfaces, human oversight and public representation.
What you need
- — Organisational BAI and technology principles
- — Model limits and human oversight
- — Privacy, representation and accountability
Paths being developed
- — BAI tool-assessment preparation
- — Labs and Build pathways in preparation
- — SEDD safety principles
Safety and participation
Safeguarding principles recordedInternal and customer data is used only for the minimum necessary purpose. A model's output is not treated as decision authority. Appeal, reversal and human approval remain visible throughout the process.
Suggested first step
Explore the relevant pathOrganisation pathway
From a tool inventory to responsible production and maintenance
When an initiative opens, a real use case is tested through a small, reversible trial. It is not expanded without human oversight and a route for appeal.
Published and developing pathways
Where can your organisation begin?
Each card shows its current status. Programmes and packages in preparation are not open demos, advisory services or approvals of fitness.
Programmes
Programmes and ways to make
Programme
In developmentLabs
Try an idea on a small scale, learn from what actually happens and decide together whether to improve it, continue or stop.
Explore the programme modelProgramme
In developmentBuild
Listen to a real need, build a small first version with the right people, test it in use and take responsibility for improving and maintaining it.
Explore the programme modelProgramme
In developmentBuilders Participation Programme
Move beyond watching: learn with others, take responsibility at your level and help sustain the work that follows.
Explore the programme modelProgramme
In developmentWAQF
Open suitable work to HAYR without leaving its rights or continuing care unattended.
Explore the programme modelProducts and services being developed
Paths shaped around real needs
Product
In developmentBAI Tool Evaluation Guide
Clear cards for the questions families, young people, schools, professions and institutions should ask before trying a digital tool, together with a small-test path.
- Six key assessment questions
- Age, data, consent and privacy card
- Small test, human review and a clear way to stop
Product
AvailableSEDD Safeguarding Guide
Practical safeguarding guidance shaped by SEDD.
- Child safeguarding
- Data and privacy
- Models and human oversight
- Representation and publication
Organisational SEDD gateway
A pilot is not expanded without four controls.
Purpose and data
Record the clear purpose, lawful basis, minimum data and deletion period.
Human oversight
Critical decisions are not delegated to a model's output; the accountable person remains visible.
Appeal and reversal
A person affected can question the decision; the process can be stopped and reversed.
Representation and accountability
Keep distinct what the model, the organisation and the person each represent.
Start here
Make one use case visible under human oversight.
The BAI Tool Assessment page sets out the first questions. The Human Oversight Guide record shows the scope of work still in preparation.
