AI guidance for teaching and research staff

This guidance is intended for all Tavistock Education and Training staff involved in teaching and research, including tutors, associate lecturers, supervisors, and professional services. It provides practical advice on using AI responsibly, ethically and securely, in line with our institutional obligations and the guidance provided by the University of Essex. Individual course programmes might also have additional guidance.

Tools such as Microsoft Copilot, ChatGPT, Gemini and Claude can help draft content, summarise information, generate ideas, analyse data and support administrative tasks. Used well, these tools can improve efficiency, enhance creativity and support decision-making. However, they also bring risks relating to accuracy, bias, data protection, confidentiality and intellectual property.

Microsoft Copilot is the Trust-approved and recommended generative AI solution for use within our organisation. Its adoption has been guided by thorough review and endorsement in line with NHS England (NHSE) guidance, ensuring that it meets the necessary standards for safety, security, and governance.

For more information, please see the NLFT Microsoft Copilot Chat guidance page. 

Our approach to AI

We encourage TET staff to explore and use AI where it can support learning, teaching, research and business operations. However, AI should enhance human expertise, not replace it. Staff remain responsible and accountable for any work created with the assistance of AI and for any decisions made using AI-generated outputs. Our use of AI is guided by the following principles: 

Integrity

AI should be used in ways that align with our organisational values, including professionalism, inclusion, equity and academic integrity.  

Human judgement and accountability

AI-generated outputs must always be reviewed, checked and evaluated by a member of staff. Human judgement remains essential, particularly where professional, academic or organisational decisions are involved.  

Transparency

Where AI has made a substantive contribution to a piece of work, staff should be open about its use. This helps maintain trust and supports good academic and professional practice.  

Security and governance

The use of AI must comply with data protection legislation, information security requirements, copyright obligations and institutional policies.  

Appropriate uses of AI

Staff may find AI useful for: 

• brainstorming ideas and generating initial drafts 
• summarising documents, reports and meeting papers 
• improving the clarity, readability or accessibility of written content 
• drafting emails, communications and reports 
• generating teaching materials and learning activities 
• supporting lesson planning and curriculum development 
• identifying themes or trends within non-sensitive datasets 
• generating alternative explanations of complex concepts 
• supporting administrative and operational workflows 

Any AI-generated content should be critically reviewed and edited before use. 

Limitations of AI

Although AI systems often produce fluent and convincing responses, they have important limitations: 

• they can generate incorrect information (‘hallucinations’) 
• they may fabricate references, quotations or citations 
• they can reproduce biases contained within their training data 
• they may provide oversimplified responses to complex topics 
• they do not understand context in the same way as humans 
• they may not reflect current policies, research or guidance 

For these reasons, AI outputs should always be verified against reliable and authoritative sources before being used in teaching, research, clinical practice or organisational decision-making.  

Data protection and confidentiality

Before using any AI tool, staff must consider the sensitivity of the information being shared. 

Do not enter:

• personal data relating to students, patients, service users or staff 
• confidential organisational information 
• unpublished research data 
• assessment submissions or student work 
• sensitive clinical information 
• copyrighted material such as journal articles 
• any information that is not already in the public domain unless the tool has been approved for such use 

Free or personal AI accounts should not be assumed to be secure. Information entered into public AI tools may be retained, processed or used for model training depending on the provider’s terms and conditions.

AI and decision-making

AI must not be the sole basis for decisions affecting students, staff or other individuals. This includes decisions relating to: 

• recruitment 
• performance management
• progression or assessment outcomes
• admissions
• fitness to practise processes
• disciplinary matters

AI can assist by providing summaries, identifying themes or drafting documents, but final decisions must always be made by appropriately qualified staff exercising professional judgement.  

AI in teaching and assessment

Staff are encouraged to help students develop AI literacy and to engage critically with AI technologies where appropriate. Increasingly, graduates will be expected to understand how to use AI responsibly and effectively in their professions.  

When designing learning activities and assessments:

• make expectations around AI use explicit
• ensure students understand acceptable and unacceptable use
• focus on learning outcomes rather than AI detection
• encourage critical evaluation of AI-generated outputs
• consider how AI may affect assessment design and authenticity

Staff should not rely on AI-detection tools to determine whether students have used AI, as such tools are known to be unreliable.

AI in research

Researchers remain fully responsible for the accuracy, integrity and ethical standards of their work when using AI tools. AI-generated content must be checked carefully and any use of AI should comply with relevant research ethics, funder requirements, data protection obligations and publication policies.  

It is essential that caution be exercised when working with confidential, sensitive or identifiable research data.

Approved tools and University of Essex guidance

For courses that are validated by the University of Essex, staff should also follow the University’s guidance on the use of AI in business operations and professional practice. Of particular importance, the University of Essex advises that:

• staff remain accountable for AI-assisted work 
• confidential or personal data must not be processed using unapproved AI systems 
• AI should support, not replace, human judgement 
• approved institutional tools should be used with appropriate organisational accounts and security protections enabled

Further support

The guidance on this page forms part of our wider Tavistock Education and Training advice on using AI:  

• AI in education 
• using AI appropriately 
• AI tools 

AI technologies continue to evolve rapidly. If you are unsure whether a particular use of AI is appropriate, secure or compliant with institutional policies, seek advice before proceeding.