Anticipate, Reflect, Engage, Act: Exploring Generative AI and Research Ethics at King's

The King’s Institute for Artificial Intelligence, through its Doing Science Well with AI programme, is working with colleagues across the university to ensure this happens in ways that are responsible, transparent and grounded in good research practice.

Generative AI (GenAI) tools are changing how research is done across disciplines, creating new possibilities for efficiency and discovery while raising difficult questions about integrity, accountability and ethics. In autumn 2025, the King’s Institute for Artificial Intelligence and King’s Research Ethics Office ran a series of GenAI and Ethics workshops to explore these questions directly with researchers and to understand what support they need from the university.

To frame this work, we drew on UK Research and Innovation’s AREA framework for responsible research and innovation: Anticipate, Reflect, Engage, Act. In this blog, we use AREA to share what we heard in the workshops, why it matters for research at King’s, and what comes next.

Anticipate

The pace of GenAI development means universities cannot afford to wait until problems emerge before responding. Ethical frameworks, guidance and support need to evolve alongside the technology so that research using GenAI remains robust, fair and trustworthy.

As a first step, the King’s Institute for Artificial Intelligence launched an expression of interest survey in late summer 2025 to understand how researchers were using GenAI tools, how confident they felt about the ethical issues, and whether they wanted further support. The survey received 275 responses from across the university, and almost all respondents indicated that they were interested in attending future workshops on AI and research.

The survey painted a picture of both enthusiasm and uncertainty. Most respondents said they were already using AI in their research (72%), and an even higher proportion (78%) planned to use it in the future, with almost half expecting to use it regularly. At the same time, 81% placed themselves in the lower to middle range in terms of confidence about navigating the ethical aspects of GenAI use.

This helped us anticipate the kinds of questions researchers would bring to the workshops: not whether to use GenAI at all, but how to use it well, how to be open about that use, and how institutional policies and processes can support them.

Reflect

The survey findings made it clear that King’s needed spaces where researchers could reflect on GenAI together, rather than tackling these questions in isolation. In response, the King’s Institute for Artificial Intelligence and King’s Research Ethics Office designed six interactive GenAI and Ethics workshops held during autumn 2025, with participants from across faculties and career stages.

Researchers were invited to discuss:​

• How GenAI is already showing up in their research
• The ethical or practical challenges they are encountering
• What guidance, support or training would help them most

Participants took part in group activities, open discussion and an individual micronarratives task, which helped surface both discipline-specific issues and cross-cutting concerns about using GenAI responsibly in research.

We piloted a novel activity using “micronarratives” – short stories co-produced by participants and a GenAI tool about how AI is being used in research. Micronarratives had previously been tested by a research team including King’s researchers, and the workshops offered a new opportunity to use them as a reflective method: researchers could describe a situation, see how the AI interpreted or extended it, and then discuss what this revealed about assumptions, risks and values.

These reflections connect local experiences at King’s with wider sector conversations on AI, science and ethics, including national work on responsible AI use in research and education. They highlight where high-level principles are helpful, and where more practical, context-specific guidance is needed for day-to-day research practice.

Engage

The workshops were designed as conversations with researchers, not one-way training sessions. Across the six events, we heard a consistent message: researchers see real potential for GenAI to enhance their work, but they want confidence that their use of these tools is ethical, transparent and aligned with good research practice.

Through group activities, open discussion and the micronarratives task, participants highlighted:​

• Clear opportunities for GenAI to support research, for example by speeding up routine tasks or suggesting alternative framings and approaches
• Barriers to adoption, including uncertainty about what good practice looks like, concerns about data protection and confidentiality, and uneven access to tools and support
• A strong desire for practical, scenario-based guidance that sits alongside existing ethics and integrity frameworks

Many researchers also emphasised the importance of culture and perception. They wanted better knowledge and shared expectations within their departments and disciplines, so that responsible uses of GenAI are recognised and supported, rather than treated as something to be hidden or as a shortcut.​

The engagement did not stop when the workshops ended. The insights and questions gathered are now feeding into ongoing conversations at King’s, and into discussions with sector partners, about what responsible GenAI use should look like at research-intensive universities.

Act

The survey and workshops were designed to inform concrete action, not just reflection. Drawing on what we learned from researchers, King’s now aims to focus on several key areas.​

First, we are working to translate workshop insights into updates for ethics review processes and guidance, so that researchers have clearer expectations when designing and approving projects that involve GenAI. This includes clarifying how to describe GenAI use in applications, what to consider around data and consent, and how to document decision-making.​

Second, we are exploring training and toolkit-style resources that address common scenarios and questions raised in the workshops. These aim to help researchers build both confidence and practical skills, with examples that reflect the diversity of research across King’s.​

Third, we will continue to engage with sector partners to share learning and shape policy, including feeding King’s experience into the development of cross-institutional principles around GenAI usage in research. Gaining insight into King’s own context puts us in a stronger position to contribute to, and benefit from, wider work on AI, science and research integrity.​

Finally, we are exploring further opportunities to use activities like the GenAI and Ethics workshops, and methods such as micronarratives, to support ongoing reflection and dialogue as GenAI tools evolve.

Our goal is to create a research environment where innovation, good research practice, and responsible AI use go hand in hand. By taking these steps, King’s aims to set a benchmark for how universities can embrace GenAI responsibly and lead the way in shaping the future of research.

If you are part of a research-intensive university and are interested in similar questions, contact ai-institute@kcl.ac.uk to learn more about our work and upcoming cross-institutional AI ethics events.

Blog image: IceMing & Digit / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/