CHAI Events

Past Events

CHAI & CRUK-Scotland Community of Practice

Globally, rates of colorectal cancer found in people under 50 – Early-Onset Colorectal Cancer (EO-CRC) – have been increasing for decades. Why this is the case, and what can be done to counteract it, was the motivation for the workshop held on the 7th of November at the Institute of Genetics and Cancer at the Western General Hospital in Edinburgh. The workshop was jointly organised by Cancer Research UK Scotland Centre (CRUK-SC) and the Causality in Healthcare AI Hub (CHAI) as part of a Community of Practice.

Colin Maclean, a data scientist from CRUK-SC, introduced the problem and the prevalence of EO-CRC among those with family history and those without, the latter comprising the majority of cases and the former featuring the steepest rise in incidence. He conveyed the complexity of the different exposures and their relationships. Hana Chockler of CHAI outlined how important it is to distinguish actual causal relationships between exposures that result in disease from mere statistical associations found in the data. Only by identifying and targeting causes can we design effective screening and prevention programmes.

With this background, the various exposures and the relationships between them over the course of lifetimes were discussed in groups. We considered how broad or precise the data on the exposures needs to be, what data may be available to inform the analysis of potential causes, and how relevant the timing of exposures is. In the final session, we mapped out potential research projects that we may want to seek funding for. A key theme of this session was the need to bring varied insight and expertise into the early stage planning of any research programme. The next stage is to consult with patient representatives before shaping a grant application.

Reflections provided by Niccolo Tempini & James Lowe

Reflections by Professor Matthew Sperrin, CHAI Co-Lead & CoP organiser

Our first in-person Community of Practice event, held in Manchester on 15 October, was a resounding success! The session brought together a diverse group of about 30 participants—including patients, clinicians, healthcare professionals, regulators, and experts in AI and statistics—to focus on the crucial topic of evidence and guidelines in groups with multiple long-term conditions (multimorbidity) and polypharmacy (people on many medications).

The first half of the session comprised a range of short talks. A highlight was Stella Finlay's moving description of the challenges faced from a patient perspective. The audience was also treated to provocative questions, such as "do we still need causal AI in the age of large language models?" posed by AI researcher Chris Yau. We gained valuable regulatory insight from Patrick Muller, who provided an overview of how NICE (the National Institute for Health and Care Excellence) incorporates real-world evidence and causal AI into their guideline development, and explained how interested researchers can connect with this process.

After a break, participants divided into breakout sessions to discuss some of the challenges in polypharmacy and multimorbidity, and how (causal) AI might be able to help. Discussions covered several topics:

• How AI can help patients engage with, and co-ordinate, health professionals involved in their care.

• Use of LLMs as a 'health coach' that helps to summarise the evidence about stopping or starting different medications.

• Identifying subgroups where a different balance of risk and benefit of particular treatments might be appropriate.

We concluded with future plans for the group - Sohan Seth explained about opportunities in CHAI and plans to write a position paper and get together again. Overall, it was a fantastic session that generated a rich exchange of interesting perspectives from all participants! We look forward to continuing this vital work.

Reflections on UKAIRS 2025 from Dr. Steph Riley  

The inaugural UK AI Research Symposium (UKAIRS) took place at Northumbria University, Newcastle from 8th – 9th September. The two days were jam packed with a variety of talks showcasing the work being undertaken as part of UKRI funded projects and programmes.  

There was a big CHAI presence at the conference, presenting our recent research across the hub.

Across the conference there was a heavy emphasis on trustworthy AI and using AI for good. Beyond the uses of AI in healthcare, it was also great to see how AI is being used to improve many aspects of society. Some of my personal highlights were using AI to detect transphobic content online and the ethical considerations when using AI in the music industry.  

The main conference was followed by an early career researcher (ECR) day on 10th September, which focused on career development for ECRs. I particularly enjoyed the session on engaging with policy in AI research, which talked about how researchers can get involved in developing and improving government policies.  

All in all, UKAIRS was a great experience to learn about different AI research taking place across the country. It was fantastic to bring the whole hub together and see everyone beyond the Teams screen!