GCR Policy Newsletter (6 March 2023)

Estimating extreme risk, interfacing science and policy, and modelling the effects of asteroid impacts

This twice-monthly newsletter highlights the latest research and news on global catastrophic risk. It looks at policy efforts around the world to reduce the risk and policy-relevant research from the field of studies.

Broken computer chips, in the style of David Hockney (produced by DALL·E 2)

---

GCR in the media

“Historically, the United States has taken a reactive and haphazard approach to preventing lab accidents and the misuse of high-risk science. A patchwork of regulations, guidance and policies exists based on the specific pathogen being researched, the type of research being conducted and the source of funding. But some research doesn’t fall under any agency, leaving an oversight vacuum. This fragmented system has not kept pace with the evolving risk landscape. There are now more powerful tools for genetic engineering, and these tools are easier to use and more widely available than ever before.” Biology Is Dangerously Outpacing Policy (New York Times)

“Despite years of efforts from professionals and researchers to quash any and all comparisons with apocalyptic science fiction and real-world artificial intelligence (AI), the threat of this technology going rogue and posing a serious threat to survival isn’t just for Hollywood movies. As crazy as it sounds, this is increasingly a threat that serious thinkers worry about. ‘AI could pose a threat to humanity's future if it has certain ingredients, which would be superhuman intelligence, extensive autonomy, some resources and novel technology,’ says Ryan Carey, a research fellow in AI safety at Oxford University’s Future of Humanity Institute.” Why risk analysts think AI now poses a serious threat to us all (ITPro)

---

Latest policy-relevant research

Estimating extreme risk in national risk assessments

The likelihood and impact assessments of five global risks - electricity supply shortage, nuclear accident, pandemic, severe space weather and volcanic outbreak - differ in the national risk assessments of nine European countries, according to a report by the Center for Security Studies (CSS). For example, the likelihood of a severe nuclear accident varies by up to 10,000 times between different assessments, and for a large volcanic outbreak by up to 5,000 times. Furthermore, the differences in the approaches between countries shows that there remains ambiguity over what should be included in a national risk assessment and that the standard 5x5 matrix can introduce distortions in risk communication and priorization. This effect can lead to the underestimation of risks with extreme impacts, such as a pandemic. (8 February 2023)

Policy comment: National risk assessments could be a powerful tool to inform governments and decision-makers about threats and hazards to the nation, including global catastrophic risks. However, methodological issues typically result in GCR being unidentified or underestimated. Many national risk assessments have short time horizons, poor risk identification processes, lacklustre expert engagement and inability to evaluate uncertainty. So complex, long-term or emerging risks tend to be ignored. A first step to fixing national risk assessments to better capture GCR would be to extend time horizons to at least ten years and prioritise risks with high degree of uncertainty (perhaps due to lack of historical record or insufficient data) regardless of estimated likelihood.

---

Interfacing science and policy

We need to reinforce more significant synergies between science and knowledge production and policy practice to ensure that research outputs are more relevant to decision-making and support more resilient governance outcomes, according to a report by a Science-Policy Interface (SPI) expert group organised by the Centre for the Study of Existential Risk (CSER). Achieving action and transformative change from science outputs is often impeded by misalignment across networks and communities of practice. This frequently restricts the advancement of both science-in-policy and policy-in-science. The complex, cascading, and system issues in the global risk context demand major transformation in mindset and policy implementation. (20 February 2023)

Policy comment: Translating the research, analysis and recommendations from the field of existential and global catastrophic risk studies into a policy context is a key blocker for advising policy-makers on these issues. Mechanisms that bridge this gap are highly useful, but only if they have strong buy-in from policy-makers and respond to concrete policy requirements. These mechanisms could also be government-led. For example, governments could develop a cross-government team from civilian and defence research and science agencies to support policy development on global catastrophic risk.

---

Modelling the effects of asteroid impacts

Given that humanity might not discover a hazardous asteroid in time for a deflection or disruption mission, civil defense mitigation strategies would also be needed, according to US Geological Survey research scientists. The detrimental effects of a medium-sized asteroid impact could last for years. The traditional phases of mitigation, relief, and recovery may need to be refined for a medium-sized impact event, or at least, planning considerations should recognize that these phases may have significant overlap or may be discontinuous. Emergency managers, resource managers and planners, and research scientists involved in mitigation and recovery efforts would likely derive significant benefit from a framework linking multiple hazard models to provide a seamless sequence of related forecasts. (6 February 2023)

Policy comment: Governments should build more sophisticated and integrated hazard and vulnerability models that account for multiple natural and space hazards as well as their potential cascading effects. According to the paper, current multi-hazard models are in their infancy and are tailored for a specific event. These improved models could better warn policy-makers and other stakeholders, such as state and local authorities, about immediate response efforts - evacuations, changes to transport corridors, emergency services. Building in the vulnerability side of the risk equation into the model would also provide more accurate assessment of long-term impacts, such as economic loss and health impacts. These assessments could inform more targeted and cost-effective recovery efforts.