Responsible AI
Last reviewed: June 2026 · Privacy policy
Responsible AI means using artificial intelligence to assist people rather than replace their judgement — with clear limits on what data a model sees, what decisions it can influence, and how its use is recorded. Journey uses AI to help skilled professionals: summarising reviews, explaining funding risk, supporting curriculum and learning-plan authoring, and surfacing what needs attention. It is never used to make high-stakes decisions on its own. These are the principles we hold AI to.
Where Journey uses AI
AI in Journey is scoped to a small set of assistive tasks:
- Drafting curriculum, learning plans and learning resources for staff to review and adapt.
- Summarising progress reviews and explaining where a learner or cohort is at risk.
- Surfacing what needs attention on per-role dashboards, so staff see the right priorities first.
What Journey's AI never does
- It does not compute or decide funding figures — those are calculated server-side against versioned DfE rule packs.
- It does not decide gateway, EPA or compliance outcomes — those are confirmed by people and rules.
- It does not receive learner personal data such as email, date of birth, National Insurance number or full address.
Human in the loop
AI output is a draft or a prompt for a person, not an automated decision. Funding, gateway, EPA and compliance outcomes are computed and confirmed by people and rules — not by a model. Staff are expected to review AI-assisted content before it informs a decision.
Privacy by design
We minimise what AI ever sees. Prompts are scrubbed of personal data — learner email, date of birth, National Insurance number and full address are never sent. Tenant-scoped helpers verify that any referenced record belongs to your organisation before a call is made, so one provider's data can never be pulled into another's prompt.
Governed & metered
Every AI call is metered and recorded in an AI usage ledger that stores compact metadata — never the raw content. Per-organisation rate and token limits apply to standard users, with controlled elevated allowances for designated senior roles. This keeps usage monitored and accountable, and lets operators see how AI is being used.
Validated & safe to fail
AI responses are validated against an expected shape before use. If a model is unavailable or returns something unexpected, the feature degrades gracefully to a deterministic result rather than failing or guessing.
Accuracy & oversight
AI can be wrong. Staff review AI-assisted content before it informs a decision, and the append-only audit log records the actions people take — so the record reflects human decisions, not model output.
Frequently asked questions
What is responsible AI in apprenticeship software?
Responsible AI means using artificial intelligence to assist people rather than replace their judgement, with clear limits on what data it sees, what decisions it can influence and how its use is recorded. In Journey, AI drafts and summarises to save staff time, but funding, gateway, EPA and compliance outcomes are always computed by rules and confirmed by people.
Does Journey use learner personal data in AI prompts?
No. Prompts are scrubbed of personal data — learner email, date of birth, National Insurance number and full address are never sent to a model. Tenant-scoped helpers verify that any record referenced in a prompt belongs to your organisation before a call is made.
Can AI make a funding or compliance decision on its own?
No. AI output is a draft or a prompt for a person. Funding figures, gateway and EPA readiness, and compliance outcomes are computed server-side against versioned rules and confirmed by staff — never decided by a model.
How is AI use kept within budget and monitored?
Every AI call is metered and recorded in an AI usage ledger that stores compact metadata — never the raw content. Per-organisation rate and token limits apply to standard users, with controlled elevated allowances for designated senior roles, so usage stays bounded, monitored and accountable.
What happens if the AI model is unavailable?
Features degrade gracefully. If a model is unavailable or returns something unexpected, the affected feature falls back to a deterministic result rather than failing or guessing, so the platform keeps working.
Journey is independent software. It is not endorsed by, affiliated with, or approved by Ofsted, the Department for Education or any other public body. References to frameworks and funding rules describe the standards Journey is built to support.
Contact
Questions about how we use AI? Email support@journeyapp.co.uk.