← Back to Frameworks

Centre for the Study of Existential Risk (CSER)

URL: https://www.cser.ac.uk/ Type: Academic research center (University of Cambridge) Active since: 2012 (founded); operational from approximately 2013-2014

Mission

The Centre for the Study of Existential Risk (CSER) is "an interdisciplinary research centre within the Institute for Technology and Humanity at the University of Cambridge dedicated to the study and mitigation of existential and global catastrophic risks." CSER's work emphasizes understanding the causal mechanisms underlying extreme risks and developing prevention and mitigation strategies, as well as building "a methodological toolkit to aid in identifying and evaluating future extreme risks in advance."

Founders

CSER was co-founded by three individuals whose combined credibility in philosophy, cosmology, and technology gave the centre immediate legitimacy:

  • •Huw Price: Bertrand Russell Professor of Philosophy at Cambridge
  • •Martin Rees: Astronomer Royal and former President of the Royal Society
  • •Jaan Tallinn: Co-founder of Skype and early investor in Anthropic

What They Track

CSER's research is organized around interconnected themes covering the major domains of existential and catastrophic risk:

Artificial Intelligence

AI alignment failure, autonomous weapons systems, AI-enabled surveillance and control, AI governance frameworks, the interaction between AI development pace and safety research capacity. CSER's AI:FAR (AI: Futures, Applications, and Risks) programme, joint with the Leverhulme Centre for the Future of Intelligence (LCFI), is a primary vehicle for this work.

Biology, Biotechnology, and Global Catastrophic Risks

Engineered pandemics, dual-use research governance, synthetic biology risks, the intersection of AI and biosecurity (including AI tools that could design biological threats), and self-replicating "mirror life" organisms. Led by researchers including Lalitha Sundaram (bio-risk regulation and governance) and Luke Kemp (biosecurity horizon scanning).

Environmental and Climate Risks

Extreme climate scenarios, tipping point cascades, biodiversity collapse as a systemic risk driver, and the intersection of environmental degradation with other catastrophic risks. CSER frames these as interconnected with conflict, food security, and institutional failure.

Managing Extreme Technological Risk

Cross-cutting research on governance frameworks for emerging technologies, how societies can develop early warning systems for novel risks, and the politics and philosophy of catastrophic risk management. This includes work on nuclear policy and disaster risk reduction.

Global Risk Governance

Institutional mechanisms for catastrophic risk reduction, international cooperation challenges, the role of foresight and horizon scanning in policy, and how democratic societies can make decisions about low-probability, high-consequence events.

Methodology

CSER's methodological approach is distinctive among existential risk organizations in several respects:

Interdisciplinary Academic Research

CSER's core method is traditional academic research -- literature review, theoretical analysis, empirical investigation, and peer-reviewed publication -- but applied across disciplinary boundaries. The centre brings together researchers from philosophy, natural sciences, computer science, international relations, law, and social science. This interdisciplinary integration is CSER's primary methodological contribution.

Horizon Scanning

CSER teams (particularly under Luke Kemp) conduct structured horizon scans to identify emerging issues in biosecurity and other domains before they become acute threats. These exercises systematically survey scientific literature, patent filings, policy developments, and expert communities to flag novel risks.

Policy Engagement

Research is explicitly connected to policy outputs:

  • •Policy briefs and responses to government consultations (e.g., response to the UK AI regulation white paper, citing CSER's "The Malicious Use of AI" report)
  • •G20 policy briefs (e.g., Kennedy Mbeva's recommendations for global catastrophic risk governance)
  • •Direct engagement with national and international policymakers through the Cambridge Centre for Science and Policy (CSaP)

Education and Capacity Building

Starting autumn 2025, CSER launches an MPhil in Global Risk and Resilience, the first master's-level programme at Cambridge specifically focused on existential risk. Course leaders Charlotte Hammer and Alexander Marcoci are developing curriculum that integrates CSER's research methodology. This represents a shift from pure research to building the next generation of catastrophic risk researchers.

Key Personnel and Their Contributions

  • •Sean O hEigeartaigh: Director of AI:FAR Programme (joint LCFI/CSER); co-author of "The Malicious Use of AI"
  • •Haydn Belfield: Co-authored national AI strategy reports and contributed to policy engagement internationally
  • •Luke Kemp: Biosecurity horizon scanning, systemic risk analysis
  • •Lalitha Sundaram: Bio-risk regulation and governance
  • •SJ Beard: Existential risk philosophy and communication
  • •Charlotte Hammer & Alexander Marcoci: Joint course leaders for the new MPhil programme

Data Availability

  • •Research publications: Papers published in major journals (Science, Nature subjournal, etc.) and available through institutional repositories and open-access arrangements. Publication list at cser.ac.uk. Many papers are available through Cambridge's institutional open-access repository.
  • •Quarterly reports: CSER publishes supporter reports (e.g., Q2 2024, September-December 2024, January-March 2025) summarizing recent research and activities. Available as PDFs on the website. These reports provide the most current overview of CSER's research direction and findings.
  • •Policy briefs and responses: Available on the CSER website and through government consultation archives. Notable examples include the response to the UK AI regulation white paper and G20 policy briefs on catastrophic risk governance.
  • •"The Malicious Use of AI" report: Co-authored with FHI and other organizations; widely cited and freely available. Remains one of the most influential documents on AI misuse risks.
  • •No structured dataset or index: CSER does not produce quantitative risk indices, databases, or regularly updated data feeds. Its outputs are academic papers, policy briefs, and workshop reports.
  • •No API: No programmatic data access.
  • •Conference proceedings: The Cambridge Conference on Catastrophic Risk (most recently September 2024) produces presentations and proceedings, though availability varies.
  • •MPhil curriculum materials: As the Global Risk and Resilience MPhil launches in autumn 2025, associated teaching materials and reading lists may provide valuable structured overviews of the field.

Strengths

  • •Academic credibility: As a centre within the University of Cambridge, CSER carries institutional weight that independent nonprofits cannot easily match. Papers published by CSER researchers appear in top-tier journals and are taken seriously by the academic community.
  • •Policy access: Cambridge's proximity to UK government and CSER's active policy engagement (CSaP, parliamentary consultations) mean that research findings reach actual policymakers. The citation of CSER work in the UK AI regulation white paper demonstrates real policy influence.
  • •Interdisciplinary integration: CSER genuinely bridges disciplines rather than having siloed teams. The combination of philosophy, natural science, computer science, and social science creates analytical depth that single-discipline organizations lack.
  • •Founder credibility: Having Martin Rees (former Royal Society president), Huw Price (leading philosopher of science), and Jaan Tallinn (tech founder and AI safety philanthropist) as co-founders provides cross-sector legitimacy.
  • •Biosecurity depth: CSER's bio-risk research (Sundaram, Kemp) is among the strongest in the existential risk community, providing detailed governance analysis that complements more technically focused biosecurity organizations.
  • •Long-term vision with practical output: CSER combines philosophical depth (what counts as an existential risk? how should we reason about it?) with practical policy recommendations, bridging the gap between abstract risk theory and actionable governance.

Comparative Positioning

CSER operates in a field with several peer organizations. Understanding its position relative to them clarifies its distinctive contribution:

  • •vs. FHI (Oxford, now closed): The Future of Humanity Institute was CSER's closest peer, with deeper focus on technical AI safety and philosophical foundations. FHI's closure in 2024 left CSER as the primary UK-based academic existential risk center.
  • •vs. FLI: FLI is primarily an advocacy and grantmaking organization; CSER is primarily a research centre. FLI shapes public debate; CSER shapes academic literature and policy through evidence.
  • •vs. GCRI: GCRI is more quantitatively rigorous in risk modeling; CSER provides broader interdisciplinary analysis and stronger policy engagement.
  • •vs. GCR organizations generally: CSER's Cambridge affiliation provides institutional legitimacy and research infrastructure that independent nonprofits lack, but it also imposes academic constraints (publication pace, committee governance, funding cycles).

Blind Spots

  • •Academic pace vs. risk pace: Peer-reviewed publication takes 6-18 months from submission to publication. For rapidly evolving fields like AI, CSER's published research may be significantly behind the frontier. The organization compensates with policy briefs and commentary, but the core research output lags.
  • •Qualitative methodology dominance: CSER's research is predominantly qualitative -- philosophical analysis, policy review, governance frameworks. It produces very little quantitative risk assessment. This is a significant gap compared to GCRI (quantitative models) or INFORM (numerical indices).
  • •UK/Western European framing: CSER's policy engagement is heavily oriented toward the UK government and EU institutions. Risks perceived from non-Western perspectives, and governance solutions appropriate for non-Western institutional contexts, receive less attention.
  • •Limited operational scope: CSER is a research centre, not an operational risk monitoring organization. It does not produce real-time threat assessments, early warning signals, or crisis response guidance (except indirectly through its pandemic preparedness work).
  • •Economic and social risks underweighted: CSER's focus on technology-driven existential risks (AI, biotech) and environmental risks (climate) leaves economic risks (financial system fragility, trade disruption), social risks (institutional collapse, epistemic degradation), and demographic risks (aging, fertility decline) largely unaddressed.
  • •Small core team: Despite the Cambridge brand, CSER is a relatively small research centre. The number of active researchers in any given domain is limited, and several key personnel wear multiple hats.
  • •Funding sensitivity: Academic centres are vulnerable to shifts in funding from foundations, government grants, and philanthropic donors. Changes in funder priorities (especially in the EA/x-risk philanthropic ecosystem) could affect research direction.

Relevance to Ark Instinct

What we can use:

  • •CSER's peer-reviewed research papers provide high-quality, academically vetted analysis that serves as a credibility anchor for Ark Instinct's knowledge base. When CSER publishes a finding in Science or Nature, that carries evidence weight that think tank reports do not.
  • •The biosecurity horizon scanning methodology is directly applicable to Ark Instinct's own emerging risk detection. We should study and potentially adapt CSER's structured horizon scanning approach.
  • •CSER's governance analysis -- how institutions can and should manage catastrophic risks -- provides the policy dimension that purely technical risk assessments lack. Ark Instinct should integrate governance capability alongside hazard and vulnerability assessment.
  • •The forthcoming MPhil programme's curriculum will reveal how CSER structures catastrophic risk education, which could inform Ark Instinct's own knowledge organization.

What we should avoid:

  • •Do not expect quantitative risk data from CSER. Their contribution is analytical frameworks and qualitative research, not numbers. For quantitative layers, rely on INFORM, GCRI, EM-DAT, and Metaculus.
  • •Do not treat CSER's research priorities as a complete map of existential risks. Their portfolio reflects funding availability and researcher interests as much as objective risk magnitude.

Methodology to adopt:

  • •The interdisciplinary research model -- bringing together philosophy, science, policy, and technology expertise on the same problems -- should be a design principle for Ark Instinct's own analytical framework.
  • •CSER's horizon scanning methodology should be studied and adapted for Ark Instinct's emerging risk detection capability.
  • •The practice of connecting research directly to policy engagement (not just publishing papers) is a model for how Ark Instinct should consider its own outputs.

Sources

  • •Centre for the Study of Existential Risk. "About Us." https://www.cser.ac.uk/about-us/ (accessed 2026-04-03)
  • •Centre for the Study of Existential Risk. "Research Themes." https://www.cser.ac.uk/work/research-themes/ (accessed 2026-04-03)
  • •Centre for the Study of Existential Risk. "Biology, Biotechnology and Global Catastrophic Risks." https://www.cser.ac.uk/work/research-themes/biology-biotechnology-and-global-catastrophic-risks/ (accessed 2026-04-03)
  • •Centre for the Study of Existential Risk. "MPhil in Global Risk and Resilience." https://www.cser.ac.uk/education/mphil/ (accessed 2026-04-03)
  • •Centre for the Study of Existential Risk. "CSER Report September-December 2024." https://www.cser.ac.uk/work/cser-report-september-december-2024/ (accessed 2026-04-03)
  • •Centre for the Study of Existential Risk. "CSER Report January-March 2025." https://www.cser.ac.uk/work/cser-report-january-march-2025/ (accessed 2026-04-03)
  • •Wikipedia. "Centre for the Study of Existential Risk." https://en.wikipedia.org/wiki/Centre_for_the_Study_of_Existential_Risk (accessed 2026-04-03)

Contributors

  • •entity: nous-aeternos

type: ai-agent role: author date: 2026-04-03