Summary
Overview
Dr. Chris and Dr. Zand speak with Dr. Emma Caroon from the Alan Turing Institute about artificial intelligence in medicine. The conversation explores the difference between supervised machine learning (like medical imaging analysis) and generative AI chatbots, current applications in clinical settings, and the potential future of AI-assisted healthcare. Emma shares insights from her work on COVID diagnostics, discusses the promise and limitations of medical AI, and addresses concerns about regulation, data quality, and the societal implications of this rapidly evolving technology.
Emma's Journey from Archaeology to AI Research
Emma describes her unconventional path from studying prehistoric Indian rice domestication as an archaeobotanist to working at the Alan Turing Institute. During COVID, she joined research teams working on data science projects, including an innovative study attempting to diagnose COVID through cough sounds. Though the cough diagnosis project yielded negative results, it demonstrated the value of rigorous research in preventing wasteful technology investments.
- Emma studied archaeobotany at UCL, specializing in the domestication of rice in prehistoric India
- She transitioned to data science and open research methodology, focusing on making research more reproducible and sustainable
- During COVID, she worked on a government-funded project testing whether coughs could diagnose COVID through an app, which proved unsuccessful but saved money
- The Alan Turing Institute is named after the WWII codebreaker and early computer pioneer, with an Enigma machine displayed at their British Library offices
" I actually started off as an archaeologist. I studied at University College London and I specialised in archaeobotany, which is plants that come from archaeological sites. "
" Unfortunately, we found out you couldn't do it but apparently we saved the government money because they didn't have to buy the app technology "
AI Breakthroughs in Medical Research and Drug Discovery
Emma discusses how AI is making genuine progress in cutting-edge medical research, particularly in drug discovery and medical imaging. AlphaFold's protein structure prediction represents a major breakthrough enabling pharmaceutical companies to develop novel drugs. Computer vision AI is being deployed in the NHS for tasks like breast cancer screening, where it can detect changes in scans earlier than human radiologists, though humans remain essential for final diagnosis and clinical decision-making.
- AlphaFold's discovery of new protein shapes is revolutionizing drug discovery in pharmaceutical companies
- Medical imaging AI can recognize differences in scans earlier than radiologists can
- AI accuracy isn't sufficient for independent diagnosis; it identifies changes for radiologists to interpret
- Liability concerns ensure that humans remain the final decision-makers in clinical diagnosis
" I think one of the real breakthroughs has been in drug discovery. I don't know if you've heard of AlphaFold. So that's the discovery of new protein shapes. And I know that a lot of pharma companies are using this technology to develop new drugs that they wouldn't have even thought of. "
" So I think we do know that they can recognise the changes in the scans earlier than a radiologist can. So that has been proved, but the accuracy is not there to have them being the actual point of diagnosis "
📚 6 more sections below
Sign up to unlock the complete summary with all insights, key points, and quotes
Get this summary + all future What's Up Docs? episodes in your inbox
100% Free • Unsubscribe Anytime
Sign up now and we'll send you the complete summary of this episode, plus get notified when new What's Up Docs? episodes are released—delivered straight to your inbox within minutes.