April HE AI community meetup - Artificial intelligence

(note: last updated V1.1 13th May 2026 – see here for the update history.)

At our April HE AI community meetup, colleagues from across the sector came together to explore emerging questions about AI in higher education. Using our usual lean coffee format, the discussion focused on AI and GDPR, student responses to AI in learning and assessment, and what an AI-shaped future might mean for universities and lecturers. This post summarises the main points shared in the discussion.

AI and GDPR

One of the most discussed topics this month focused on AI and GDPR, particularly the use of tools such as Microsoft Copilot with enterprise protection. There was a strong sense of uncertainty around what is considered safe and appropriate, especially when working with personal or sensitive data.

There was detailed discussion about how Microsoft Copilot with enterprise protection handles prompts, uploaded files and web search functionality in practice. Several participants raised questions about what happens to information once it is entered into the tool, particularly when working with personal or sensitive data

Some of these questions are explored further in Jisc’s blog Microsoft Copilot and Data Protection: An Update.

One area discussed in particular was how Copilot uses web search to enhance responses. The concern was that personal information might leave the Microsoft 365 tenancy via a Bing search. To be clear, the prompt itself is protected under enterprise data protection. The question is what gets sent to Bing when a search is triggered.

Sometimes a Copilot query will trigger a Bing search to enhance the response. In this case, key phrases from the prompt may be used to carry out the search. The full prompt is not normally sent directly to Bing unless the prompt is very short. Uploaded files are not sent directly to Bing and remain within the organisation’s Microsoft 365 environment.

However, prompts may still contain personal or sensitive information. For example, a prompt such as “look for ways I can provide wellbeing recommendations for John Smith informed by the latest best practice published by the NHS” still contains identifiable and potentially sensitive information. This prompt is also likely to trigger a Bing search, as the NHS information is unlikely to be held within your Microsoft 365 tenancy. As an example, in our test it triggered a search for “NHS workplace wellbeing best practice”, so the personal information was removed. We cannot guarantee this happens in every instance, though, so it makes sense to apply data minimisation principles.

The prompt did not need to include the staff member’s name, in the same way it would not need to if you carried out the search yourself. Staff should therefore avoid entering unnecessary personal or sensitive information into prompts, even when using a tool with enterprise protection, especially where the task involves linking to external sources of information.

The discussion also highlighted the need for clearer staff guidance and better understanding of how enterprise AI tools handle prompts, web search functionality and sensitive information in practice.

Student responses to AI use

A significant part of the discussion focused on how students are responding to AI in learning and assessment. Staff shared a wide range of experiences from their own institutions, highlighting the diversity of student responses.

Some participants described lower engagement in sessions where AI was introduced, while others noted hesitation around using AI in learning activities. In other contexts, students were more open, particularly where AI was introduced gradually or framed as part of developing digital and academic skills.

In many cases, these responses were linked to ethical and practical considerations. Students raised questions about environmental impact, bias, and the wider implications of AI use. In some subjects, particularly in the arts, this extended to questions of identity and professional practice.

This prompted discussion about student autonomy and how to support participation when AI is part of an activity or assessment. Several participants suggested that offering alternative routes may help ensure all students can engage in ways that align with their perspectives. This was framed both as an inclusion consideration and a practical aspect of assessment design.

Differences across disciplines were also noted. Participants reported varying responses across disciplines with some noting more openness in STEM contexts. Students in these areas were often described as more open to using AI, while in arts and humanities there was greater variation in response. Some participants observed that attitudes can shift when AI is introduced through an AI literacy approach, rather than positioned as a required tool.

Trust and ethics were recurring themes throughout the discussion. Participants also reflected on how media coverage and wider public narratives may be influencing student perceptions.

A key question was how institutions should respond. Should students be required to engage with AI, or should the focus be on helping them understand it and make informed choices? While there was no single view, many participants leaned towards building AI literacy and critical awareness as a constructive starting point.

Universities and lecturers in an AI world

We also explored what AI might mean for the role of the university and the lecturer.

Participants shared examples of AI-generated teaching content, including avatars and tools such as HeyGen, as well as AI-generated video formats used in education. This prompted discussion about where AI-generated content is starting to appear in teaching, and what that might mean for more traditional approaches.

Student perspectives were again a central theme. Several participants reported that some students are less comfortable with AI being used in teaching, while others described more mixed feedback. For example, some students felt that AI-supported marking could offer greater consistency or reduce bias.

Overall, this points to a balance that institutions are still working through, between potential benefits and students’ expectations of a human-centred learning experience.

The discussion also touched on how AI is framed, including narratives around inevitability. Participants noted a range of views, from enthusiasm to scepticism, reflecting the wider public conversation.

One area of interest was the use of AI across both learning and assessment processes. This raised questions about how best to design teaching and assessment in ways that remain meaningful, transparent and grounded in academic judgement.

While no single conclusion was reached, there was a shared recognition that AI is beginning to shape aspects of teaching practice. At the same time, many felt that the lecturer’s role continues to centre on trust, interaction and human connection.

Key takeaways

Discussions around Microsoft Copilot highlighted the need for clearer staff understanding of how prompts, web search functionality and enterprise protection interact in practice, particularly when working with sensitive information.
Student perspectives on AI are varied and often shaped by trust, ethics and context. Many participants highlighted the importance of supporting AI literacy so students can make informed decisions.
AI may offer additional ways to create and present teaching content. The group explored how this can be balanced with expectations for human involvement in teaching, feedback and assessment.

What’s next

Thank you to everyone who joined and contributed. Our next HE AI community meetup takes place on the 19^th of May. If you’d like to join the next session, please Join the AI in higher education community meetup list

Links shared during the call

Change log

V1.1 13 May 2026

Added AI and GDPR section.

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