Over the last year, I have spent time with 462 staff in colleges and universities across the UK. These were in teaching and learning roles. I wanted to hear their views on artificial intelligence: how they are using it, what excites them, and what concerns they carry.
These conversations happened during a period of rapid change. Generative AI has moved from being peripheral to becoming central in education debates. People are now asking big questions about pedagogy, assessment, workload, learner equity, and even the purpose of teaching itself.
One thing is clear to me: staff are not adapting from a level playing field. Colleges and universities might share the same intent but the resources, infrastructure, and support available to them are very different. This is creating a two-speed system. College staff often told me they are relying on free tools, workarounds, or even their own personal accounts just to keep up. Unless there is targeted investment and joined-up planning, these divides will deepen further.
This blog is a companion to our Student Perceptions of AI 2025 report. Together they show both sides of the story. The staff perspective is essential: it captures how people are responding in real time. Staff are experimenting, adapting, questioning, and sometimes resisting. These findings also complement Jisc’s 2025 Leadership Survey. That survey showed most institutions now have AI policies in place but far fewer have embedded them in everyday practice or invested in the development staff need to feel confident. In other words, there is a gap between policy and practice, and staff are the ones navigating it.
What I am hearing from staff
Staff are pragmatic and creative. They are using AI to save time, support planning, personalise materials, and enhance engagement. Most are self-taught and experimenting ahead of institutional guidance.
Assessment systems are struggling to keep pace. Staff want to redesign tasks to reward process and authenticity but are slowed down by governance structures and awarding bodies.
Confidence is uneven. Some are highly engaged while others feel excluded or anxious. Without structured support these gaps will widen.
Policies exist, but often stay on paper. Staff say they are poorly communicated or not reflected in daily practice. Leadership needs to bridge the gap between intent and action.
Ethical worries are increasing: bias, transparency, sustainability, and the tension of being told to use AI professionally while restricting its use by students.
Professional roles are shifting. Staff question the relevance of what they teach, especially in creative and digital fields. Some see erosion of traditional skills, whilst others are pragmatic about change.
Policy is lagging behind practice. Staff already use AI daily, but guidance does not reflect this. Misalignment between staff and student policies is confusing.
Colleges face particular pressure. Lack of funding, limited access to paid tools, and a reliance on free versions is creating inequity.
The wider education landscape is fragmented. Learners face inconsistent approaches as they move between school, college, university, and the workplace.
How staff are using AI
The practical examples I heard were wide-ranging and include:
Supporting teaching, learning, and assessment
- Lesson planning: drafting schemes of work, unit plans, and learning outcomes. Creating ideas for new topics or fresh takes on old ones.
- Interactive materials: designing Mad Libs style activities for grammar, storytelling exercises, descriptive writing supported by AI-generated imagery, and creative prompts. Using Copilot and MagicSchool for presentation and resource generation.
- Assessment design: generating exam questions, rephrasing to improve clarity, and stress-testing multiple-choice items. Drafting rubrics and assessment criteria.
- Marking and feedback: using AI to mark binary answers, draft feedback, and check alignment with grading criteria. Some are trialling whole assignment analysis using AI to suggest improvements.
Supporting accessibility and inclusion
- Creating scaffolds and summary sheets for learners with additional needs or limited digital access.
- Translating materials into languages such as Ukrainian and Bengali, sometimes in real time.
- Simplifying texts for screen readers or learners who need accessible formats. Some staff are exploring AI use for visually impaired learners.
Supporting administration
- Drafting emails in different tones, from formal to marketing.
- Redesigning schedules and timetables.
- Writing policy, action plans, or formal responses.
- Creating Excel formulas and generating reports from data.
How staff are embedding AI in student learning
Many staff make their own use of AI visible to students. They want transparency and to model responsible practice. Examples include:
- Running workshops on research skills, misinformation, and prompt design.
- Teaching digital literacy: how to ask better questions and critique AI outputs.
- Supporting writing: helping students brainstorm, refine, and redraft.
- Language support: giving English as additional Language (EAL) learners tools to improve vocabulary and clarity.
In creative fields, the use is even more striking. Fashion tutors are guiding students to use AI for trend research and concept development. Music staff are using it to generate drum patterns and MIDI sequences to support composition. Media staff are experimenting with scripts, descriptive visuals, and narrative structures.
Experimentation and learning
Many staff are still exploring. They are testing Copilot, ChatGPT, and other tools to understand limitations and possibilities. Common activities include:
- Playing with prompts to see what improves outputs.
- Sharing knowledge in informal groups or peer-led sessions.
- Taking part in internal workshops where offered.
- Experimenting with AI detectors to learn how they work and where they fail.
- Creating AI hubs to share practice and collaborate.
Concerns staff are raising
Student use
Some staff worry that students are leaning too heavily on AI, skipping essential thinking and learning. In trades and practical subjects, this is especially concerning: you cannot rely on AI on a building site, but some learners are doing just that.
Others worry that students copy and paste multiple-choice questions into AI tools instead of engaging with the exercise. In creative subjects, there are fears about originality being lost: music staff spoke about students trying to pass off AI-generated work as their own. Across the board, staff worry about over-reliance, passivity, and loss of confidence in personal ability.
Academic integrity
There are widespread concerns about cheating. AI detectors are unreliable. Some students leave giveaway phrases like “here is the answer to your question” in their work. At the same time, staff themselves have been wrongly flagged by detection tools. This leaves many uneasy. Oral questioning is often used as a check. Some institutions send suspected cases to panels, but staff say professional judgement remains essential. Many also spoke about the moral tension of being asked to use AI in their jobs while telling students not to.
Assessment
Staff are frustrated by the slow pace of assessment reform. Many said current tasks are not fit for purpose in the age of AI. They want to design tasks that draw out critical thinking and authentic voice, but the processes take too long. At one university, staff said it took six months to change an assignment. Without greater flexibility, AI misuse will increase because outdated tasks invite workarounds.
Staff confidence
Confidence levels are uneven. Many staff are self-taught. Without clear guidance, they are unsure what is allowed, which tools are approved, or how to use them responsibly. Even where support is available, time pressures prevent staff from engaging. This fuels anxiety. Digital gaps are widening between staff and students, and between staff themselves. Some mature learners and staff struggle with the basics, while others are advanced.
Some also expressed frustration that expectations for AI use are growing, but without extra time or resource. They feel pressure is rising faster than capacity.
Ethical concerns
Bias in outputs is a big issue. One example staff shared was image tools generating racially biased depictions of criminals. Others worry about hidden human labour in AI supply chains. Transparency is another recurring theme: staff question how to ensure honesty when AI content can be indistinguishable from human work. Some raised the environmental impact of large AI systems.
Job security and curriculum relevance
Some staff fear AI could replace parts of their role. Creative and teaching staff feel especially vulnerable. Others take a pragmatic view, seeing AI as a tool rather than a threat, but all agree the landscape is shifting.
There are deep questions about curriculum. Are we still teaching the right things? Skills like composition, coding, and analysis are changing as AI takes on more tasks. Some staff fear what they teach is already becoming obsolete. Others argue the challenge is not preventing change but helping learners adapt to new competencies.
Professional identity
Many staff feel a loss of craft. Designing original resources, tailoring lessons from experience, and handwriting feedback are all being automated. This is disorienting. Without clear leadership, staff are left to make personal decisions about how much to rely on AI. They want guidance that goes beyond reassurance: clarity about priorities, role expectations, and the future value of teaching itself.
What staff say they need
- Clear guidance: simple, accessible policies for staff and students. Which tools are approved? What counts as acceptable use? How can assessments be adapted? Without clarity, staff are cautious or inconsistent.
- Development: not just one-off sessions but ongoing training, reflection, and hands-on practice. Peer learning and AI champions are especially valued. Staff want pedagogy combined with practical prompting skills and ethical awareness.
- Time and digital support: space to experiment without fear. A baseline offer to ensure no one is left behind.
- Supportive culture and leadership: staff want senior teams to back experimentation, IT to provide practical support, and colleagues to champion good practice. Confidence grows when people feel they are learning together.
Jisc’s Leadership Survey 2025 showed that only 44% of FE and 37% of HE institutions have delivered staff development on AI. Most have policies, but policy without practice is not enough. Staff are calling for consistency across institutions to reduce confusion and policy drift.
Curriculum agility is another urgent need. Current systems take months to update assignment briefs. External awarding organisations and professional bodies are not moving quickly enough either. Staff feel trapped between outdated tasks and rapidly evolving tools.
Admissions also came up. University staff worry about personal statements being written with AI. Questions of fairness and authenticity remain unanswered. Sector-wide guidance is needed.
Structural imbalance
The AI transition is magnifying old inequalities.
- Resources: colleges have fewer licences, less training capacity, and limited funding. Staff often rely on free versions or personal accounts. This is creating a two-speed system.
- Transitions: skills and expectations vary widely between schools, colleges, and universities. Learners often have to relearn or abandon AI practices at each stage.
- Employers: staff want clearer signals from industry about the skills that matter. Without this, curriculum planning remains speculative.
My conclusion
AI is already here and already shaping teaching. Staff are engaging pragmatically and creatively, often without formal support. Institutions have taken first steps with policies and working groups but there is a gap between ambition and implementation.
Staff concerns are real: fairness, relevance, workload, and professional identity. If we want AI to enhance rather than erode education, institutions must act decisively. That means aligning policy with practice, investing properly in staff, and creating cultures that support experimentation.
AI is not something on the horizon. It is already in classrooms, offices, and assessment processes. The challenge now is whether the sector can keep up.
Where we go from here
After listening to staff across the UK these are the main actions I’d suggest:
Staff need clear and practical guidance. They told me they want short, accessible policies they can actually use. Guidance that spells out which tools are approved, what counts as acceptable use, and how assessments can be adapted without breaching specifications. It cannot sit hidden away in long documents. It needs to be updated regularly as tools evolve. And it must be consistent across staff and students, otherwise trust breaks down.
Align Policy: We should accept that staff are already using AI every day. Policies should reflect that reality. Currently, there is a gap between official statements and classroom practice. That gap makes staff nervous and undermines confidence. If staff are to model responsible use, they need clarity and alignment, not mixed messages.
Invest in Staff Development: One-off workshops are not enough. Staff want ongoing, embedded support that is realistic about workloads. They want time to try things out, reflect, and share practice. Peer learning matters here: many trust colleagues who are experimenting alongside them more than formal trainers. AI champions can help, but only if institutions back them properly. Development must also connect practice with pedagogy. Prompting skills on their own are not enough: staff want to know how AI can support lesson design, assessment, and ethical teaching.
Prioritise Assessment Redesign: This is the area where tension is sharpest. Staff want tasks that reward process, critical thinking, and authenticity. Many realise assessments need to reflect skills that matter in a world where AI exists. But internal processes are too slow, and external awarding and regulatory bodies are not moving quickly either. Institutions need to make it easier to adapt assessments quickly at module level. Sector-wide, there must be pressure on external bodies to modernise.
Lead a Culture Shift: Staff need to feel safe experimenting without fear of getting it wrong. They want leadership that makes space for curiosity, discussion, and even failure. The danger is that we default to compliance, where people are afraid to try. We need the opposite: a culture of openness, honesty, and support. Visible endorsement from senior leaders helps, as does practical backing from IT teams and encouragement from colleagues.
Strengthen Collaboration: Right now, learners are experiencing different messages in schools, colleges, universities, and workplaces. That fragmentation undermines trust. We need consistency so learners can build on skills as they progress rather than having to unlearn them. Collaboration is also vital to address inequality. Colleges are at risk of falling behind because they cannot afford the same tools as universities. Without targeted investment, further education staff and learners will be left behind. Employers need to join this conversation too. Staff are being asked to align teaching with workplace needs, but without clear signals from industry about which skills matter most, planning feels speculative. Stronger partnerships will help curriculum and assessment reflect real-world demands.
With thanks to:
BEST Network[i]
Blackburn College
Coleg Cymoedd
Edgehill University
Scotland Focus Group[ii]
South Devon College
Ulster University
University of Bedfordshire
University of Cambridge
University of Lancaster
University of Northampton
[i] The BEST network brings together seven UK Business Schools—Adam Smith, Alliance Manchester, Birmingham, King’s, Lancaster, Liverpool, and Warwick—to support collaboration and dissemination of outputs related to the Scholarship of Teaching & Learning (SoTL).
[ii] The Scotland focus group included staff from a wide range of institutions: Abertay, Edinburgh Napier, Glasgow Caledonian, Heriot-Watt, Queen Margaret, Robert Gordon, Open University, University of Aberdeen, University of Edinburgh, University of Glasgow, University of Dundee, University of St Andrews, University of Stirling, University of Strathclyde, and the University of the Highlands and Islands (UHI).
Glossary of AI tools
Staff across colleges and universities are drawing on a wide range of AI tools to save time, personalise learning, create new types of content, and adapt to growing demands. This glossary reflects the tools mentioned throughout the report. It offers a snapshot of the tools shaping AI practice on the ground.
AI Adlibs: Used by staff to generate interactive, fill-in-the-blank stories and grammar exercises. Popular in creative writing and language learning. aiadlibs.com
Adobe Firefly: AI-powered creative tool for generating images, editing photos, and supporting visual content creation in digital arts and media. Adobe Firefly
Canva: Design platform with built-in AI tools for creating engaging visual resources, presentations, and educational materials. canva.com
ChatGPT: Widely used across teaching and admin. Staff use it for lesson planning, drafting emails, writing support, assessment design, marking, generating resources, study aids, and experimentation. chat.openai.com
Claude: An advanced conversational AI used for nuanced writing tasks such as essay structuring, feedback drafting, and reflective writing. Some staff prefer its tone and depth. claude.ai
Copilot: Microsoft’s AI assistant embedded in Office tools. Used for planning, adapting resources, scaffolding content, generating presentations, and drafting feedback. copilot.microsoft.com
Excel (AI-enabled): Spreadsheet tool enhanced with AI for generating formulas, automating reports, analysing data, and supporting admin tasks. Microsoft Excel
Free AI Detectors: Used experimentally by staff to check whether content was AI-generated. Reliability is limited, but they help staff explore detection boundaries. GPTZero
Google Gemini: Google’s AI chatbot used for summarising content, generating explanations, and comparing answers. Helpful for lesson preparation and student support. Gemini
MagicSchool: An education-focused AI tool for generating lesson plans, slide decks, and classroom resources tailored to learner needs. magicschool.ai
Midjourney: AI image generator used in creative and digital subjects such as media, design, and storytelling. midjourney.com
Pictory: AI video creation tool that turns written content into engaging videos for media and presentation purposes. pictory.ai
Runway: AI video and image editing tool used in creative subjects to generate visuals, animations, and video projects. runwayml.com
TeacherMatic: AI platform for educators. Used to create quizzes, multiple-choice questions, rubrics, and tailored teaching content. teachermatic.com
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