Global and existential risks
Global and existential risks is a barrier to human progress within the societal factors. The evidence for it, and the methods proposed to address it, follow.
Ord (2020) argues that humanity underinvests in reducing risks such as pandemics, nuclear war, climate change, and emerging technologies. His probability estimates are his own judgments, not measurements.
Evidence
- COVID-19 deaths. WHO estimated 14.83 million excess deaths worldwide in 2020 and 2021. That is 2.74 times the 5.42 million deaths reported as caused by COVID-19. Only about half of countries had complete death data, so many estimates are modeled (Msemburi et al., 2023).
- Pandemic preparedness. A model estimated expected global losses from future pandemics at over $700 billion a year. It found that spending $60 billion up front and $5 billion a year on vaccine production capacity would yield an expected net gain of more than $500 billion (Glennerster, Snyder, and Tan, 2022).
- Uncertainty. A University of Leeds policy brief argues that claims about the return on investment in preparedness rest on unreliable assumptions. So the size of the underinvestment is uncertain.
What can be done
Actions rated on the strength-rating scale (A strong and replicated, B solid but limited, C weak or debated, D contested or failed), applied to the specific claim made.
- Coordinated global eradication of a disease, smallpox (B). WHO launched an intensified eradication plan in 1967. The last known natural case was in Somalia in 1977. WHO declared smallpox eradicated in 1980. Smallpox is the only human infectious disease to be eradicated (WHO, n.d.).
- Ring vaccination to stop outbreaks (B). In Guinea, contacts of Ebola cases, and their contacts, were randomized by ring to immediate or delayed vaccination. The trial later continued without randomization in Guinea and Sierra Leone. There were no cases from day 10 among those vaccinated immediately, versus 16 cases in delayed rings. That is an efficacy of 100% (95% CI 68.9 to 100.0) (Henao-Restrepo et al., 2017). Randomization was stopped early and the trial was open-label.
- Nonproliferation treaties, the NPT (C). After accounting for which states chose to join, ratifying the Nuclear Nonproliferation Treaty was robustly associated with a lower chance of pursuing nuclear weapons from 1970 to 2000 (Fuhrmann and Lupu, 2016). The size of the effect could not be confirmed.
- Limit: weapons bans without verification, the Biological Weapons Convention (B). The convention entered into force on March 26, 1975 and has 187 parties. But the convention has no standing inspection regime. A proposed compliance protocol failed in 2001. The convention's support unit has three permanent staff. The Soviet Union maintained an enormous offensive biological weapons program after ratifying it (Arms Control Association, 2024).
- Limit: binding health rules without enforcement, the International Health Regulations (B). WHO convened a committee to review how the regulations worked during COVID-19. Like earlier review committees, it found three problems. Too many countries still lacked the required public health capacities. There was no mechanism for monitoring key provisions. And WHO had not challenged any country that reported additional travel measures, although by August 2020 all countries had imposed some. The committee also found that IHR core capacity scores alone were not a good predictor of pandemic response (WHO, 2021).
- Not shown to work: national preparedness scores as a guide to pandemic performance (C). The Global Health Security Index and WHO's Joint External Evaluation scores were strongly correlated with each other. But these scores did not predict how quickly countries detected their first COVID-19 case or their COVID-19 deaths. The share of people aged 65 and over did (Haider et al., 2020).
Proposed and experimental methods
Methods that are proposed, under trial, approved in some places, or tried and then failed. Each shows a stage label and an evidence rating. A stage label shows how far a method has progressed, not whether it works. The stage labels are explained on the societal factor page.
- Kinetic deflection of asteroids, DART (Early trial, B). NASA's Double Asteroid Redirection Test deliberately crashed a spacecraft into Dimorphos, the small moon of the asteroid Didymos. It was the first full-scale test of pushing an asteroid off course by hitting it (Daly et al., 2023). The impact shortened Dimorphos's orbit by 33.0 minutes (plus or minus 1.0 minute). That was far more than the roughly 7 minutes expected from the spacecraft's momentum alone, because debris thrown off by the impact added push (Thomas et al., 2023). It is one test on one small asteroid.
- WHO Pandemic Agreement (Approved but not scaled, B). The World Health Assembly adopted the agreement on May 20, 2025, after a committee vote of 124 in favor, none against, and 11 abstentions. The agreement includes a planned pathogen access and benefit-sharing system. Under that system, participating manufacturers would give WHO rapid access to 20% of their real-time production of pandemic products. The agreement opens for signature only after an annex on that system is adopted, and it enters into force after 60 ratifications (WHO, 2025). The United States left WHO on January 22, 2026. That was one year after an executive order of January 20, 2025 began its withdrawal. The United States also stopped taking part in WHO committees, governance bodies, and technical working groups (U.S. Department of Health and Human Services, 2026). Negotiations on the annex missed the May 2026 World Health Assembly. They were extended to May 2027 or an earlier special session, so the agreement had not opened for signature as of October 2026 (WHO, 2026).
- 100 Days Mission for pandemic vaccines (Theoretical, C). The Coalition for Epidemic Preparedness Innovations (CEPI) aims to have safe vaccines ready for initial authorization and large-scale manufacturing within 100 days of identifying a new pandemic threat. For COVID-19 vaccines, this took 326 days. CEPI cites modeling that says meeting the target could have prevented 8.3 million excess deaths (CEPI, n.d.). The target has not yet been met for a new pathogen.
- Far-UVC light to disinfect indoor air (Early trial, C). Far-UVC lamps give off 222 nm ultraviolet light. This light kills airborne microbes and is thought to be relatively safe for skin and eyes. In a room-sized chamber, five filtered lamps at the exposure limits allowed for 8 hours cut airborne Staphylococcus aureus by 98.4% (Eadie et al., 2022). In a conference room, one lamp did not significantly change ozone or fine particles. But four lamps raised ozone by 5 to 10 ppb and particle counts by 16% (Narouei et al., 2025). No trial measuring infections in people was found.
- Metagenomic and traveler-based pathogen surveillance (Early trial, C). Untargeted sequencing of wastewater can in principle detect any known or new pathogen. A modelling study of four sequencing datasets found that, after controlling for local infection rates, the share of reads from a given virus varied by orders of magnitude across studies. The same study found that, in one study, a respiratory virus enrichment panel lowered the estimated yearly cost of a system from $7.87 million to $287,000. The system was one able to detect a SARS-CoV-2-like pathogen before 0.01% cumulative incidence. But the panel was possibly reducing sensitivity to unexpected pathogens (Grimm et al., 2025). A CDC program tested pooled nasal swabs from arriving international travelers at four U.S. airports from September 2021 to January 2022. It reported the first U.S. Omicron BA.2 and the first BA.3 in North America (Wegrzyn et al., 2023). No system has yet been shown to give early warning of a new pathogen.
- Government AI safety and security institutes (Approved but not scaled, C). Governments have set up institutes to test advanced AI models for dangerous capabilities. On February 14, 2025, the UK renamed its AI Safety Institute the AI Security Institute. The renamed institute was focused on risks such as AI misuse for chemical and biological weapons, cyber-attacks, and crime. The UK said it would not focus on bias or freedom of speech (UK Government, 2025). In June 2025 the U.S. Commerce Department announced that the U.S. AI Safety Institute would become the Center for AI Standards and Innovation. The center stayed within NIST as industry's main government contact for testing commercial AI systems (NIST, 2025). As of October 2026, NIST's page names it the Center for Advancing Innovation and Standards for Super Intelligence. The page says its evaluations focus on demonstrable risks such as cybersecurity, biosecurity, and chemical weapons (NIST, n.d.). No evaluation of whether such institutes reduce risk was found.
- Reversed: U.S.-Russia strategic arms limits, New START (Failed or reversed, B). New START capped each side at 1,550 deployed strategic warheads and 700 deployed missiles and heavy bombers. Russia suspended implementation in February 2023. The United States did not respond to a September 2025 Russian proposal to keep observing the central limits for a year. The treaty expired on February 5, 2026 (Arms Control Association, 2026).
Sources cited on this page
- Glennerster, R., Snyder, C. M., & Tan, B. J. (2022). Calculating the costs and benefits of advance preparations for future pandemics. NBER Working Paper 30565. Published 2023 in IMF Economic Review, 71(3), 611-648. DOI C Limited
- Ord, T. (2020). The Precipice. London: Bloomsbury. New York: Hachette. C Limited
- Msemburi, W., Karlinsky, A., Knutson, V., Aleshin-Guendel, S., Chatterji, S., & Wakefield, J. (2023). The WHO estimates of excess mortality associated with the COVID-19 pandemic. Nature, 613, 130-137. DOI B Moderate
- Daly, R. T., Ernst, C. M., Barnouin, O. S., Chabot, N. L., Rivkin, A. S., Cheng, A. F., et al. (2023). Successful kinetic impact into an asteroid for planetary defence. Nature, 616(7957), 443-447. link B Moderate
- Eadie, E., Hiwar, W., Fletcher, L., Tidswell, E., O'Mahoney, P., Buonanno, M., et al. (2022). Far-UVC (222 nm) efficiently inactivates an airborne pathogen in a room-sized chamber. Scientific Reports, 12(1), 4373. link C Limited
- Narouei, F. H., Tang, Z., Wang, S. I., Hashmi, R. H., Welch, D., Sethuraman, S., et al. (2025). Effects of germicidal far-UVC on ozone and particulate matter in a conference room. PLOS ONE, 20(8), e0328224. link C Limited
- Thomas, C. A., Naidu, S. P., Scheirich, P., Moskovitz, N. A., Pravec, P., Chesley, S. R., et al. (2023). Orbital period change of Dimorphos due to the DART kinetic impact. Nature, 616(7957), 448-451. link B Moderate
- Fuhrmann, M., & Lupu, Y. (2016). Do arms control treaties work? Assessing the effectiveness of the Nuclear Nonproliferation Treaty. International Studies Quarterly, 60(3), 530-539. link C Limited
- Haider, N., Yavlinsky, A., Chang, Y.-M., Hasan, M. N., Benfield, C., Osman, A. Y., et al. (2020). The Global Health Security index and Joint External Evaluation score for health preparedness are not correlated with countries' COVID-19 detection response time and mortality outcome. Epidemiology and Infection, 148, e210. link C Limited
- Henao-Restrepo, A. M., Camacho, A., Longini, I. M., Watson, C. H., Edmunds, W. J., Egger, M., et al. (2017). Efficacy and effectiveness of an rVSV-vectored vaccine in preventing Ebola virus disease: Final results from the Guinea ring vaccination, open-label, cluster-randomised trial (Ebola Ça Suffit!). Lancet, 389(10068), 505-518. link B Moderate
- Grimm, S. L., Kaufman, J. T., Rice, D. P., Whittaker, C., Bradshaw, W. J., & McLaren, M. R. (2025). Inferring the sensitivity of wastewater metagenomic sequencing for early detection of viruses: A statistical modelling study. The Lancet Microbe, 6(11), 101187. link C Limited
- Wegrzyn, R. D., Appiah, G. D., Morfino, R., Milford, S. R., Walker, A. T., Ernst, E. T., et al. (2023). Early detection of severe acute respiratory syndrome coronavirus 2 variants using traveler-based genomic surveillance at 4 US airports, September 2021-January 2022. Clinical Infectious Diseases, 76(3), e540-e543. link C Limited
Every source for this factor is listed on the societal factor page.