Discoveries without understanding: Trusting AI research
Claire Malone returns to the Ri to explore whether we can trust AI generated scientific results if we don't know the full picture.
Artificial intelligence (AI) is becoming increasingly embedded in scientific research - analysing data, proposing hypotheses, and in some cases, even writing papers. We are relying on systems that may lead us to answers faster and more efficiently, but what happens when we don't understand precisely how those answers were reached? Can science, which has long valued explainability and transparency, embrace a collaborator that often produces results without revealing how they were reached?
By considering how today’s most advanced generative AI systems — such as large language models — work, Claire Malone will explore how tools like ChatGPT generate predictions not by understanding concepts but by calculating what is statistically most likely to come next in any given context, examining how this is accelerating scientific discovery.
From researchers at CERN using AI models to simulate collisions, detect anomalies, and accelerate data analysis at experiments like ATLAS, to AlphaFold - developed by Google DeepMind - which has solved the long-standing problem in biology of predicting a protein’s 3D structure from its amino acid sequence.
Despite their impressive results, the use of these AI tools raises some profound philosophical questions: is AI merely extending the power of human curiosity - or is it reshaping it? Can an algorithm that finds patterns really discover something new? Or does AI need to remain a tool worked in tandem with human scientists, with conscious judgement, creativity and doubt a vital complement to machine speed?
And finally, if discoveries could come from systems we do not fully understand, what does it mean to ‘have trust in science’?
By booking to attend events at the Royal Institution, you confirm that you have read and agree to the Ri's event terms and conditions.
Event image by Luke Jones via Unsplash
Claire Malone returns to the Ri to explore whether we can trust AI generated scientific results if we don't know the full picture.
Artificial intelligence (AI) is becoming increasingly embedded in scientific research - analysing data, proposing hypotheses, and in some cases, even writing papers. We are relying on systems that may lead us to answers faster and more efficiently, but what happens when we don't understand precisely how those answers were reached? Can science, which has long valued explainability and transparency, embrace a collaborator that often produces results without revealing how they were reached?
By considering how today’s most advanced generative AI systems — such as large language models — work, Claire Malone will explore how tools like ChatGPT generate predictions not by understanding concepts but by calculating what is statistically most likely to come next in any given context, examining how this is accelerating scientific discovery.
From researchers at CERN using AI models to simulate collisions, detect anomalies, and accelerate data analysis at experiments like ATLAS, to AlphaFold - developed by Google DeepMind - which has solved the long-standing problem in biology of predicting a protein’s 3D structure from its amino acid sequence.
Despite their impressive results, the use of these AI tools raises some profound philosophical questions: is AI merely extending the power of human curiosity - or is it reshaping it? Can an algorithm that finds patterns really discover something new? Or does AI need to remain a tool worked in tandem with human scientists, with conscious judgement, creativity and doubt a vital complement to machine speed?
And finally, if discoveries could come from systems we do not fully understand, what does it mean to ‘have trust in science’?
By booking to attend events at the Royal Institution, you confirm that you have read and agree to the Ri's event terms and conditions.
Event image by Luke Jones via Unsplash
Good to know
Highlights
- 1 hour 30 minutes
- In-person
Refund Policy
Location
The Royal Institution
21 Albemarle Street
London W1S 4BS
How would you like to get there?
