Blog by Hasini De Silva (pictured)
I’m Hasini De Silva, a Master’s Bioinformatics student passionate about the intersection of AI and biology, especially in fields of synthetic biology, protein design and drug discovery. I’m currently working on my dissertation, which focuses on engineering and optimising the redox properties of a de novo membrane protein under Associate Professor Paul Curnow. Through this project, I’ve developed a deep curiosity about how machine learning can be used not just to understand biology, but to reprogram it with precision and intent.
I’m always looking for opportunities to work on projects that bring biology and computation together. That’s why I was so excited to attend two recent AI in Health workshops hosted by the Elizabeth Blackwell Institute, a hub for interdisciplinary health research at the University of Bristol, supporting collaboration across disciplines to drive innovation and tackle complex health challenges. The first workshop focused on Socio-Digital Transformation, Ethics and Governance in healthcare. The second explored Genomics, Protein Design and Drug Discovery. Each offered something rare, a space to explore cutting-edge science and engage in cross-disciplinary dialogue that was as critical as it was inspiring.
Artificial Intelligence in Health: socio-digital transformation, ethics and governance worksop, 6 May 2025
AI in health is not just about algorithms, it’s about systems
The first workshop opened with a clear message, AI in health does not succeed on technical performance alone. Its impact depends on context, on the systems into which it is introduced, and the people it is meant to serve.
We heard from ethicists, researchers, software engineers and policy experts who explored what it means to deploy AI responsibly within healthcare. These were not abstract concerns, they were grounded in real-world clinical and institutional challenges.
Three ideas resonated strongly with me:
1. Context is essential
Dr. Emanuele Ratti introduced the concept of interactional AI ethics. He emphasised that AI must be judged not only by accuracy or efficiency, but by how well it works in the specific settings where it is used. A model that performs well in development can still fail if it disrupts patient care or undermines trust in a hospital or GP practice.
2. Sustainability is part of governance
Professor Ruzanna Chitchyan raised the important issue of environmental and cognitive sustainability. As AI systems scale, so do their costs, from energy use to maintenance demands. Responsible AI must factor in these costs from the beginning, especially in resource-constrained healthcare environments.
3. Trust is an ecosystem, not a feature
Professor Jonathan Ives reminded us that trust is not something that can simply be added onto a system. It is built through legal accountability, professional standards and cultural readiness. In clinical settings, trust in AI depends as much on policy and communication as it does on interface design or technical reliability.
One of the most meaningful parts of this workshop was how interdisciplinary the conversations were. Between sessions, I had the chance to speak with clinicians, engineers, ethicists and researchers. Everyone brought questions, not just answers, and there was a real willingness to listen and learn across disciplines.
Artificial Intelligence in Health: genomics, protein design and drug discovery workshop, 22 May 2025
From decoding biology to designing it
The second workshop turned toward the molecular frontier, showing how AI is being used not just to understand life, but to redesign it.
I was particularly excited by the discussions on programmable biology. Researchers are moving beyond protein prediction to protein design, focusing on building molecules that can carry out specific, functional tasks.
Professor Dek Woolfson, Prof. Ross Anderson and Assoc. Prof. Paul Curnow presented work on de novo protein design, where the goal is not just to predict how proteins fold, but to engineer them to switch, catalyse and transport. This is the future of function-first biology, powered by computation.
Professor Lucia Marucci demonstrated how AI and control theory can be used to create cybergenetic systems, where living cells are steered in real time using microfluidics and feedback control. This is biology as computation, happening inside cells, not just on a screen.
Dr Daniel D’Andrea introduced spatial biology tools that combine transcriptomics and histology to create actionable maps of disease. These maps go beyond data layers, providing clinical insight that is grounded in real biological context.
Using high-performance computing infrastructure like Isambard-AI, researchers are now accelerating workflows that once took days. From protein simulation to real-time inference, AI is helping bring lab-scale design closer to clinical relevance.
For me, this was incredibly energising. As someone working at the intersection of machine learning and protein engineering, it was exciting to see the field moving with such purpose and speed.
Looking ahead
These workshops offered much more than updates on cutting-edge research. They were a reminder that AI in health is not just a technical advancement, but a systems-level challenge that brings together engineering, biology, ethics and governance.
I left feeling both inspired and grateful. Inspired by researchers who are designing biology with purpose, and by ethicists and policy leaders who are working to ensure that innovation remains responsible, equitable and grounded in real-world impact.
Huge thanks to the Elizabeth Blackwell Institute for creating space for these important conversations. I’m grateful to have been part of them and look forward to staying engaged with this vibrant and growing community.
Watch the talks from the AI in Health: Socio-digital transformation, ethics and governance workshop.
Watch the talks from the AI in Health: Genomics, protein design and drug discovery workshop.
Explore our AI in Health workshop series 2024/25 via our YouTube playlists.