Since obtaining my PhD in Philosophy from the London School of Economics, I have worked primarily at Oxford University's Future of Humanity Institute (2015-16) and Cambridge's Centre for the Study of Existential Risk (2016-25). My research combines my expertise in philosophy with my experience of foresight and the developing transdisciplinary field of Existential Risk Studies to explore and interpret large-scale technological, societal, and environmental transformations and the risk they pose for humanity.
My current focus is AI safety and governance. As a PIBBSS fellow, I am developing a novel taxonomy of agency in order to better distinguish the kinds of agency human and AI systems exhibit and how they interact, to make issues such as meaningful control, the evaluation of human/AI coalitions, and the handover of decision-making to AI systems conceptually tractable and practically governable. This continues a method I have applied throughout my career, of taking problems that resist clean definition and building the taxonomies, frameworks, and foresight tools to make them legible and manageable, drawing on other safety-critical domains such as biosecurity, nuclear energy, and disaster studies. The four groups of work below trace how I apply that method.
My current focus is AI safety and governance. As a PIBBSS fellow, I am developing a novel taxonomy of agency in order to better distinguish the kinds of agency human and AI systems exhibit and how they interact, to make issues such as meaningful control, the evaluation of human/AI coalitions, and the handover of decision-making to AI systems conceptually tractable and practically governable. This continues a method I have applied throughout my career, of taking problems that resist clean definition and building the taxonomies, frameworks, and foresight tools to make them legible and manageable, drawing on other safety-critical domains such as biosecurity, nuclear energy, and disaster studies. The four groups of work below trace how I apply that method.
Firstly, I have studied the drivers of existential risk and how to manage them, focusing on issues of systemic risk and complexity. As a PIBBSS fellow, I am currently developing sociotechnical perspectives on AI safety, centred on understanding how humans, AI, and other agents interact to form complex systems that might, or might not, support human flourishing. I have also explored this issue in previous work, including a report on the Malicious Use of Artificial Intelligence and an essay on how educational institutions should prepare for AI's transformative potential. Previous work has explored other risk domains, including interdisciplinary studies of climate change's contribution to existential risk (in Futures); how to manage the potential risks and benefits of solar geoengineering, with Gideon Futerman; and Japan's 3/11 triple disaster (the world's costliest technological disaster) and it's aftermath, with Julius Weitzdorfer. My most significant contribution to our understanding of risk has been developing ways to study existential risk holistically (rather than partitioning it into separate 'risks') and developing governance strategies that address our existential vulnerabilities and exposures, as I set out in my book Existential Hope.
Secondly, I have developed and applied novel foresight tools. I have extensive experience with Horizon Scanning and helped develop TERRA, an early machine learning-based horizon scanning tool in 2018. I have also led a participatory horizon scanning exercise, using a modified version of the Delphi technique, on AI and Higher Education (for Microsoft's AI Economy Institute), and participated in similar exercises focusing on food security, climate change, and public health. With Rick Davies, Lara Mani, and Tom Hobson, I helped develop the ParEvo participatory foresight tool for scenario development and risk exploration, while my study into possible futures for the Science of Global Risk was the first peer-reviewed academic publication using the technique. This tool allows for more creative and divergent perspectives on possible futures to be synthesised and uses an ethnographic approach that makes it well suited to engaging a wide variety of stakeholders. I am an editorial board member of the journal Futures, supporting its publication of world-leading existential risk research and the use of quantitative modelling, and I have advised the UK Parliament and civil servants on risk foresight.
Thirdly, I have produced foundational work on the methodologies and norms of Existential Risk Studies. I surveyed and evaluated the methods currently used in the field, with Tom Rowe and James Fox, and called for greater rigour and reflexiveness in their application, even when evidence is limited and issues are urgent. I assessed seven contemporary challenges facing it with Lalitha Sundaram and Matthijs Maas and proposed different ways to address them. I wrote the field's first academic history, with Emile Torres, arguing that it has experienced three waves of development, each building on the last while broadening its scope of perspectives and approaches. I published an evidence-based report on how to improve diversity and inclusion across the existential risk community, with Suzy Levy. Finally, I led a foresight study to explore possible futures for the field, using the ParEvo technique. Drawing on this work, I have co-edited two books on Existential Risk Studies published by Open Book: The Era of Global Risk: An Introduction to Existential Risk Studies (with Martin Rees, Catherine Richards, and Clarissa Rios-Rojas) and An Anthology of Global Risk (with Tom Hobson). This work helped me to launch the world's first masters programme in Existential Risk, Cambridge University's MPhil in Global Risk and Resilience.
|
|
|
Finally, I have contributed to our understanding of the ethics of future generations and human extinction, an extension of my PhD research, mainly in collaboration with my colleague Patrick Kaczmarek. Across two papers (in Utilitas and Argumenta) and two book chapters, we developed social contract-based arguments against human extinction: looking at our obligations to past and present people, to future generations themselves, and to humanity as a whole. We offer this as an alternative to dominant, utilitarian-based ethical paradigms for thinking about human extinction. Dr Kaczmarek and I have also published papers arguing in favour of a positive time preference for wellbeing, known as social discounting, and that we should give priority to preventing near-term disasters over seeking to alter the long-term trajectory of humanity (in The Monist) and exploring Derek Parfit's Theory X. I have also published my own response to Derek Parfit's Repugnant Conclusion and axiological paradoxes it creates. I have drawn upon this work in advocating for the representation of future generations in policymaking.
The multifaceted nature of my work reflects both the breadth of my interests and my desire to work collaboratively with others to gain a bigger picture of complex problems. I started out doing ethics, but the more I thought about how to respond to the difficult choices we face, the more I realized we didn't properly understand them yet, and that we needed new tools for doing so. My ongoing research is also inately bound up in both my approach to teaching and my policy, media, and communications work.
My Google Scholar profile is here.
The multifaceted nature of my work reflects both the breadth of my interests and my desire to work collaboratively with others to gain a bigger picture of complex problems. I started out doing ethics, but the more I thought about how to respond to the difficult choices we face, the more I realized we didn't properly understand them yet, and that we needed new tools for doing so. My ongoing research is also inately bound up in both my approach to teaching and my policy, media, and communications work.
My Google Scholar profile is here.
Group portrait for the Centre for the Study of Existential Risk in September 2024