AI policy and practice in further education

Further education colleges face a different set of AI-related questions to secondary schools. Vocational and technical qualifications, mixed adult and 16-18 cohorts, employer partnerships, and a distinct funding and inspection landscape all shape how AI policy should be approached in FE. This guidance sets out the main areas colleges need to consider, distinct from a school-focused approach.

This is general guidance, not a template policy or legal advice. Colleges should check current guidance from awarding bodies, the Education and Skills Funding Agency (ESFA), and Ofsted's FE-specific inspection materials directly, since requirements in this area continue to develop.

Why FE needs its own approach, not a school policy adapted

A policy written for a secondary school context often assumes a single curriculum model, a single age group, and GCSE/A-level style assessment. FE colleges typically need to account for:

• A much wider range of qualification types (vocational, technical, apprenticeships, adult skills, access courses), each with different awarding body rules.
• Learners across a wide age range, from 16-year-olds to adult returners, with different expectations of independence and different safeguarding considerations.
• Close integration with employers, including work placements and employer-set assessments.
• Funding streams and compliance requirements specific to the sector.
• A separate FE-specific Ofsted inspection toolkit, rather than the schools toolkit.

Because of this, an AI policy adapted line-by-line from a school template is likely to miss areas that matter most in a college context.

Vocational course integration

AI tools intersect with vocational and technical education in ways that don't map neatly onto academic subjects:

Where AI genuinely supports vocational delivery:

• Generating practice scenarios or case studies for vocational assessment practice (for example, simulated workplace situations).
• Supporting the production of clear, step-by-step process documentation, which mirrors the kind of documentation learners will encounter in many vocational settings.
• Helping staff keep pace with fast-changing industry practice by summarising updates in a given sector, which staff should verify before using.
• Supporting English and maths functional skills delivery for learners who need additional scaffolding.

Where more care is needed:

• Practical, competency-based assessment (for example, observed skills in construction, health and social care, or engineering) generally can't be meaningfully replaced or assessed by AI tools, and any AI use here should be limited to supporting materials rather than the assessment itself.
• Where a vocational qualification is linked to an external professional or regulatory body (for example in health, care, or safety-critical trades), any AI-related teaching content should reflect that body's current standards, checked directly rather than relied on from AI output.
• Employer-facing assessment (such as work placement sign-off) should remain a human judgement, with AI, if used at all, limited to generating supporting documentation.

A useful principle for curriculum teams: AI can support the production and delivery of vocational content, but rarely the underlying competency judgement itself.

Assessment integrity in a vocational context

Assessment integrity concerns in FE differ from schools in a few specific ways:

• Many FE qualifications are assessed continuously through coursework, portfolios, or practical observation, rather than a single terminal exam, which changes where AI misuse is most likely to occur and how it's detected.
• Awarding bodies vary in their current position on AI use in assessed work, and this differs by qualification type, so subject teams should check the relevant awarding body's guidance for each qualification they deliver, rather than assuming a single college-wide position covers every course.
• Apprenticeship end-point assessment has its own integrity requirements, set by the relevant assessment organisation, and any AI policy should be checked against those specific requirements.
• Functional skills and GCSE resit qualifications, often delivered within FE colleges, follow the same exam board rules as those qualifications carry in a school context.

Given how much this varies by qualification and awarding body, colleges are generally better served by a college-wide statement of principles (declaration expectations, what counts as misuse) with detailed, course-specific rules sitting underneath it and reviewed by each curriculum team against their own awarding body's current guidance.

Staff CPD needs

FE staff often come from industry backgrounds rather than teacher training routes, which affects what AI-related CPD needs to cover:

• Basic AI literacy and safe use should not be assumed, particularly for staff who moved into teaching directly from industry and may not have covered this in initial teacher training.
• CPD should cover both classroom/pedagogical use (using AI to support planning, differentiation, and resource creation) and sector-specific use (how AI is showing up in the industries learners are training for).
• Given how quickly AI tools and expectations are changing, a one-off training session is unlikely to be sufficient; an ongoing, updated approach works better.
• Existing sector bodies, including Jisc's National Centre for AI, produce FE-specific staff guidance and resources that are updated periodically and can usefully sit alongside a college's own CPD programme, rather than colleges building everything from scratch.
• CPD should also cover the limits of AI tools relevant to specific vocational areas (for example, where AI output in a technical or regulated subject may be outdated or incorrect, and needs to be checked against current standards).

Funding and compliance context

FE colleges operate under funding and compliance requirements that don't apply to schools in the same way, and these have some bearing on AI adoption decisions:

• Data handling requirements tied to funding audits (for example, learner records used for ESFA funding claims) mean that any AI tool touching learner data used in funding calculations needs the same data protection scrutiny as any other system handling that data, and colleges should treat this as a compliance-relevant use case, not just a safeguarding one.
• Apprenticeship funding and end-point assessment involve external assessment organisations and employer co-investment, adding another party whose expectations around AI use may need to be checked.
• Where AI tools are procured using public funding, colleges should apply the same procurement and value-for-money scrutiny expected of other funded purchases.
• Ofsted inspects FE and skills providers using a separate toolkit to the schools toolkit, so college leaders should refer to the FE-specific materials, not the schools framework, when considering how AI use might be evidenced at inspection.

Building a college AI policy that reflects this context

Rather than adapting a school policy, colleges are generally better served starting from:

• A short statement of principles covering safe, ethical, and effective AI use across the college, similar in spirit to sector-level principles already published by bodies such as Jisc and the Association of Colleges.
• Course-level guidance sitting underneath this, reviewed against each awarding body's or assessment organisation's current position, since this varies by qualification.
• Staff CPD that's ongoing rather than one-off, covering both general AI literacy and sector-specific application.
• Clear links to existing data protection and safeguarding processes, particularly where learner data feeds into funding claims.
• A review point tied to the college's own inspection cycle and to any changes in awarding body or funding body guidance.

A sector still finding its footing

AI guidance for FE is still developing, and current sector-level resources make clear that best practice is still emerging across the sector, with colleges encouraged to share what works. Building a policy that's specific to vocational delivery, assessment structures, and the FE funding and compliance landscape, rather than one borrowed from a school context, is likely to serve colleges better as this guidance continues to develop.

EssingtonITS works with colleges on AI governance and safe adoption.