Staff Guidance for FE

Context

Since the release of Jisc’s original Staff Guidance for Further Education in early 2024, the pace of change in Artificial Intelligence (AI) has accelerated significantly. New tools and features are now embedded across everyday platforms such as Microsoft, Google Classroom, Virtual Learning Environments (VLEs), and everyday workflows. Colleges are beginning to see both the opportunities and challenges that come with widespread AI adoption.

In response, this updated guidance reflects the rapidly changing landscape. It builds on our own Jisc internal guidance but has been adapted for colleges.  

This updated guidance builds on Jisc’s internal staff guidance and has been adapted specifically for colleges. It aims to help staff use AI safely, ethically, and confidently in their daily work, making best use of existing tools to enhance efficiency, reduce workload, and improve the learner experience.

The following section contains example guidance for staff, which may be adopted or adapted to align with your college’s own policies and context.

Guidance for staff

Building AI Literacy and Skills

We believe that equipping staff with the skills and knowledge to use AI well should be at the heart of every college’s approach. By helping staff to build confidence and proficiency in using AI, we aim to empower you to use AI effectively and responsibly, to identify new opportunities, and to create further efficiencies and benefits across the organisation. The goal is to foster a culture of continuous learning and innovation. 

Participate in the college’s training programme to help staff use AI effectively, responsibly, and critically. Training will cover key topics including: 

When working with AI, you’ll often be using data. You should be familiar with the college’s data protection policies and secure working practices, including guidance on data handling, information classification, and cyber security. 

Using AI for Teaching and Learning

AI’s potential to improve efficiency is more than just a promise; it is already proven in practice. For example, our Teachermatic pilot showed a saving of around three hours a week for teaching staff. 

AI can help staff streamline processes, automate tasks, and reduce workload, allowing more time to focus on more valuable work. It can also support more effective decision-making, though final responsibility should always lie with a human. Human oversight remains key to helping the college manage risk, and in some cases, human involvement is legally required. 

One of the challenges of getting started with AI is knowing exactly how it can help you. To help with this, we are sharing a range of practical, everyday examples. These examples demonstrate the current breadth of AI’s application and underscore its potential to streamline routine tasks, increase efficiency, and ultimately free up staff to focus on more complex responsibilities. If you aren’t sure how to use these prompts, then booking onto Jisc’s AI literacy training will help. 

Title	Description	Example Prompt
Creating learning resources	Suggests activities, discussion prompts or examples aligned to qualification outcomes or curriculum standards.	“Create three classroom activities for Level 2 Health and Social Care on communication skills. Include: learning objective, 10–15 minute activity steps, materials, differentiation for Entry Level readers, one quick formative check per activity, and a short debrief question. Align to City & Guilds outcomes. Write in plain English. Add a short note on how to adapt for small groups vs whole class.”
Creating lesson plans	Helps plan structured sessions with timings, outcomes and resources.	“Draft a 90-minute lesson plan for Level 3 Digital Production on ethical use of AI. Include: objectives mapped to the qualification spec, timed agenda, resources, teacher script bullets for key explanations, two checks for understanding, one practical task, success criteria, and extension tasks. Add accessibility notes: font sizes, alt text expectations for slides, and guidance for learners using screen readers.”
Differentiating learning materials	Adjusts text or tasks for different literacy levels or learning needs.	“Rewrite the following paragraph on renewable energy for three reading levels: Entry Level, Level 1, Level 2. Keep key terms accurate. Use short sentences and everyday words at lower levels. After each version, add two comprehension questions and one vocabulary check. Source text: [paste].”
Generating assessment questions	Creates formative or summative questions to test understanding.	“Create a bank of 20 assessment items on workshop health and safety for Level 1 learners: 10 multiple choice, 5 short answer, 5 scenario questions. Provide answer keys and brief rationales. Tag each item with topic and difficulty. Ensure language is accessible and avoids trick questions. Flag any items that may need local policy alignment.”
Providing constructive feedback	Suggests formative feedback to support reflective learning.	“Using the reflection below, produce feedback comments that are supportive, specific and action-focused. Structure as: What went well, Even better if, Next steps. Keep tone encouraging and avoid rewriting the learner’s work. Limit to 120 words. Reflection: [paste].”
Writing learning support plans	Structures support plans and strategies for individual learners.	“Draft a concise learning support plan for a learner with dyslexia. Sections: strengths, challenges, classroom adjustments, assessment arrangements, assistive tech recommendations, communication preferences, review date. Keep to one page. Use bullet points in plain English. Inputs from tutor notes: [paste].”
Analysing learner data	Identifies patterns from attendance or progress data to inform interventions.	“Analyse this attendance spreadsheet for Level 3 Engineering. Produce: a 6-bullet summary of patterns, a list of at-risk learners with reasons, and three practical interventions we can try next week. Include a short note on data limitations. File: [attach].”
Summarising and producing reports	Summarises long documents and drafts concise internal reports.	“Summarise this Ofsted report for senior leaders in 10 bullets: key strengths, priorities, immediate actions. Then draft a one-page exec summary for staff. Use neutral, evidence-based language. Add a table of three actions, owners and timelines. Source: [paste or attach].”
Drafting job descriptions	Produces consistent role descriptions for teaching and support roles.	“Write a job description for a Learning Resource Centre Assistant. Include purpose, key responsibilities, essential and desirable criteria, typical hours, safeguarding statement and EDI commitment. Use clear, inclusive language and avoid jargon. Output as a clean, scannable layout.”
Supporting administrative communication	Drafts professional internal messages, updates or reminders.	“Draft an email to all curriculum staff summarising today’s team meeting. Include three decisions, assigned actions with owners and deadlines, and the date of the next review. Keep to 180 words max. Add a short postscript with where to find the minutes and who to contact with questions.”

Collaborating and sharing

We want staff to collaborate and share how they are using AI across the college. Teams that share openly are finding they move faster, learn more, and avoid repeating the same work. Sharing what works well, and also what doesn’t, is helpful. 

This approach promotes knowledge exchange and helps staff learn from each other’s experiences, building confidence and consistency across departments. By sharing insights, challenges, and lessons learned, we can identify best practices, improve efficiency, and foster a culture of responsible experimentation. 

If you have examples, prompts, or workflows that others might benefit from, share them through your college’s internal community channels or AI working group. If these don’t yet exist, consider setting one up to encourage cross-college collaboration and keep discussions active as practice evolves. 

Guidelines for Tools Incorporating AI and Data Usage

	Microsoft Copilot for Education Microsoft Copilot is integrated across Microsoft 365 applications such as Word, PowerPoint and Teams. It is covered by the same enterprise data protection and privacy standards as other Microsoft tools. Inputs and outputs are not used to train public AI models, and all data remains within your organisation’s Microsoft environment. Copilot can assist with drafting text, creating summaries, and analysing data. Staff should still review outputs for accuracy and appropriateness before sharing them. Always use your institutional Microsoft account to ensure full data protection.
	Google Gemini for Education Google Gemini is available through Google Workspace for Education and offers AI-powered support across Docs, Slides, and Gmail. When provided as a core service under institutional control, data is protected within the college’s Google environment and is not used for public model training. Staff should use Gemini to enhance productivity and creativity but must always check AI-generated content for factual accuracy and bias. Do not upload confidential or personal data outside the managed Workspace environment.
	ChatGPT for Education ChatGPT can generate, summarise, and explain text or support lesson planning and communication. Institutional licences such as ChatGPT for Education provide stronger privacy and security than the public version, ensuring data is not used to train public models. If using the free or personal version, turn off chat history and avoid entering confidential or identifiable information. Always review and adapt outputs to suit your learners and ensure accuracy before use.

Websites with AI Features

Many websites also include AI features. This includes everyday tools such as Google Search, as well as web-based AI tools such as ChatGPT. As with any website not provided as part of the college’s approved list, you should never input any information that would be classified as private or confidential, personal (that identifies an individual), intellectual property (of the college or another person or organisation) and/or otherwise protected. 

Understanding AI Chat Tool Risks

AI chat tools are trained to answer in a human-like way. Because of this it’s easy to assume that these tools learn in a human-like way too from the data we enter. That’s not really the case. Most of their training data is scraped from the internet or obtained through contracts with information providers, and they don’t learn as they go. They are re-trained or fine-tuned by the developer as they create new versions of the underlying model – sometimes every few weeks, and often less. The chances of them repeating details from a single document they are trained on are very low. 

This doesn’t mean we shouldn’t be careful with what we enter into AI tools and websites. Although they might not ‘learn’ from the data, unless the college has a robust contract in place with the provider of the tool, we don’t know what they are doing with that data or whether it is kept secure. For example, the database of the high-profile chatbot DeepSeek wasn’t secured properly, meaning anyone could potentially see all the chat history. Of course, this issue isn’t unique to AI systems, but because they answer in a human-like way, are easy to use, and produce helpful outputs, this may make us more tempted to input sensitive, confidential or proprietary data into them. Only do this with tools provided by the college through the IT department and in accordance with the college’s policies and procedures. 

Safeguarding and Age Restrictions

When using any tools with learners, including AI, staff must consider age limits and consent requirements. These exist to comply with UK GDPR, the Children’s Code, and the Online Safety Act. 

Approved education tools such as Microsoft Copilot for Education (13+) and Google Gemini for Education can be used safely through college-managed accounts, as they include enterprise-level data protection and do not use learner data for model training. 

Some tools, including ChatGPT, Perplexity, and Grammarly, require parental or guardian consent for under-18s, depending on how they are deployed. Always check the provider’s current terms of use before introducing a tool in class. 

Staff should ensure that only age-appropriate, approved systems are used with learners and that consent is in place where required. 

For more detail, and up-to-date guidance see Jisc’s Age Restrictions and Consent to Use Generative AI (August 2025).  

Copyright

We are often asked about copyright and AI. Generative AI models rely on the use of large datasets for their training, which may include copyright works. If these works are used without permission there is a risk that this will infringe the copyright of third parties, both when the work is copied by a developer to train the AI tool, and when outputs are created by the AI tool. Third party copyright may also be infringed if we include copyright works without permission when inputting data into the AI tool. 

You should never use AI to intentionally produce content that infringes on the copyright of others, for example, by copying their work. AI systems will sometimes produce material or outputs that are very similar to existing copyright works due to the data on which they have been trained. 

Microsoft provides an indemnity that protects colleges against claims in this area when college-provided Microsoft services are used, so you are free to use these tools to produce or assist with content in accordance with applicable policies and procedures. 

Google provides enterprise-level security and contractual data protection through its Workspace for Education terms. Colleges can use Gemini to create or assist with content in line with institutional data protection and acceptable use policies.   

OpenAI provides enhanced privacy and data protection through ChatGPT for Education and enterprise agreements. Colleges can use ChatGPT to create or assist with content in line with institutional data protection and acceptable use policies. 

This is a complex area, and the government is still consulting on its approach. The college is keeping updated on changes in government policy, guidance, or legislation that might require a change of approach to AI use. 

Accuracy 

Generative AI tools can and will make mistakes. As a user, it’s your responsibility to ensure outputs are accurate and reliable. This is particularly important in areas where AI is used in teaching materials, reports, or may influence decision-making, or where AI outputs will be published externally to the college. 

Aligning with Data Ethics

AI use in colleges should align with key data ethics principles that promote trust, fairness, and accountability: 

Privacy and Personal Data

These principles also form part of the college’s data protection and privacy obligations. We must, of course, ensure we follow the college’s policy on privacy when we use any technology, and AI is no different. If you have any questions relating to AI and privacy, please refer to the college’s Data Protection Officer or digital learning team for advice. 

Commitment to Sustainability and Environment

There is growing concern about the environmental impact of AI. It also has broader implications across the UN’s Sustainable Development Goals, including economic, social, and environmental dimensions. 

As with anything, we should use AI mindfully and carefully. As standards emerge that make it easier to assess and report the precise environmental impact of AI tools, we will explore how to make these available. However, these standards are not yet in place. 

You should strive to make sure that AI-generated content is free from bias to promote fairness, accuracy, and inclusivity in our decisions and communications, and to avoid creating biased content that could negatively impact individuals, our college and our stakeholders. 

Accessibility and Inclusivity

AI tools can be powerful aids for improving accessibility and supporting inclusive learning. They can help staff produce descriptive text (alt-tags), adjust the tone or reading level of learning materials, or deliver content in new ways such as through transcription or text-to-speech services. AI can also help adapt resources to meet the needs of learners who use assistive technology or require additional support. 

However, as with any digital tool, staff should always check that the materials created or adapted using AI meet accessibility standards and reflect the diverse needs of all learners. 

It’s also important to be aware that AI-generated content may include bias. For example, images created using AI can show gender or racial bias, and text may reflect cultural assumptions from the data used to train these systems. Always review and, where necessary, edit outputs to ensure they are accurate, fair, and inclusive before sharing with learners. 

Final Thoughts

AI offers significant opportunities to improve efficiency, personalise learning, and support inclusive practice across further education. Realising these benefits depends on how it is used — thoughtfully, transparently, and with clear human oversight. 

Jisc will continue to develop practical guidance, training, and resources to help colleges make the best use of AI while protecting staff, learners, and data. 

 

---

Find out more by visiting our Artificial Intelligence page to explore publications and resources, learn more about our communities and sign up for our AI Literacy training.

For regular updates from the team sign up to our mailing list.

Get in touch with the team directly at AI@jisc.ac.uk