The opening days of the Russian invasion of Ukraine brought a flurry of dramatic changes to the world. Now, three weeks into the war, conditions on the ground have somewhat stabilized. Further major changes remain very much possible, but meanwhile, there is now an opportunity to reflect on what has happened and what that means for the world moving forward. This post provides some reflections oriented mainly, but not exclusively, for a global catastrophic risk audience, with some emphasis on nuclear weapons.
This one matters. The war itself poses significant risks and it has a variety of important implications for international affairs. People who work on global catastrophic risk and related issues will benefit from understanding this. One might wonder how much time to spend studying the war. While each person should decide based on their own circumstances, I can say that studying it is not a waste of time.
Whether to try getting involved to help on the immediate situation is a more complex matter. It’s already getting massive attention and effort, including from people who are highly trained to work on this sort of situation. I would be cautious about trying to get involved based on a limited understanding of the issues. For example, understanding the long-term moral importance of global catastrophic risk is a weak basis for making specific policy recommendations. There’s a lot of nuance that goes into international security policy that also needs to be accounted for.
For my part, I have focused on activities that are clearly beneficial and that are not already getting significant attention. On several occasions, I’ve gotten an idea for an activity only to find, upon closer inspection, that it’s a bad idea or that it’s already been done several times over by people more qualified than myself. This is despite the fact that do have some modest background on international security issues, especially nuclear weapons. I am a generalist who sometimes works on nuclear weapons, whereas there are communities of people who specialize on the topic. A healthy dose of humility has helped me avoid making some inappropriate contributions.
More detailed discussion is below. Links to suggested readings are at the end.
Nuclear War Risk
The possibility that nuclear weapons could be used in the ongoing war is a matter of grave concern. Several knowledgeable observers have postulated that the conflict could escalate to nuclear war. From a risk perspective, this raises the questions of what the probability of nuclear war is and how severe it would be. Some estimates of the risk of the Ukraine war escalating to nuclear war have been made by people who are active in geopolitical forecasting. As I discuss in a recent article on nuclear war risk analysis, attempts to quantify the ongoing risk raise significant methodological challenges.
First, nuclear war risk is an inherently difficult risk to quantify. It depends on highly complex and geopolitical factors. A lot of the relevant information is not publicly available. Nuclear war is more-or-less unprecedented; the only precedent, World War II, occurred under circumstances that no longer exist.
Second, the ongoing Ukraine war involves constantly changing circumstances. A risk estimate that is good one day may be bad another day. As an illustration, in the run-up to the war, Robert de Neufville’s estimate of the probability that Russia attacks Ukraine went from 65% to 25% in around one to three days circa Feb 12-15 due to apparent lower tensions. de Neufville is a superforecaster and a former colleague of mine at the Global Catastrophic Risk Institute, where we co-authored a GCRI paper on the probability of nuclear war. The fact that his probability estimates changed so dramatically doesn’t mean that de Neufville was wrong or a bad forecaster (my impression is that he is quite good at this), just that events can change fast.
For my part, I was a lot more worried about nuclear war in the opening days of the Ukraine war than I am now. At first, events were changing so quickly that there appeared to be significant risk of them spiraling out of control. Since then, conditions have stabilized into a state that is tragic for Ukraine but safer for the rest of the world. However, the events are still playing out and there is still plenty of chance that the events could take a turn for the worse, suggesting there is currently an elevated nuclear war risk relative to normal, calm conditions.
How the war ends could have significant implications for the ongoing risk of nuclear war and likewise the importance of nuclear war within the overall portfolio of global catastrophic risk topics. If the war is resolved without coming close to nuclear war, it raises the question of whether nuclear war may not be a significant risk. The Ukraine war is such a rare, extreme crisis that if this doesn’t come close to nuclear war, then perhaps nuclear war is exceptionally improbable. Alternatively, perhaps this war may avoid coming close to nuclear war specifically because certain efforts were made to avoid coming close, in which case perhaps similar efforts may be important to retain within the overall portfolio of work on global catastrophic risk. Or, perhaps the war will come close to nuclear war—indeed, perhaps it already has and information about it is not yet public—or perhaps it will escalate to nuclear war. We should hope not, but this is not over yet.
The Geopolitical Landscape
The ongoing geopolitical changes are large and are likely to have major implications for a wide range of issues. The Western democracies and their global allies have found renewed sense of purpose and assertiveness even while they wisely avoid military escalation with Russia. This bodes well for their future prospects, and for international governance more generally.
Russia is a more difficult case. It is important to resist the temptation to view the country with contempt and indulge in schadenfreude at their military and economic struggles. Russia the country did not invade Ukraine; the decision was made by a narrow set of government leaders, in particular Putin. Public opinion surveys in Russia are hard to come by; what I’ve seen shows mixed support. Regardless, the international community should not want an isolated and impoverished Russia. That would be bad for the Russian people and could result in the Russian military leaning more heavily on its nuclear weapons. It could further risk a dangerous sense of humiliation and resentment among Russians that may motivate them to regroup and pursue renewed conquest. The humiliation of Germany following World War I is a case in point of what to avoid. Arnold Schwarzenegger’s message to the Russian people strikes me as a good example of how to reinforce their sense of dignity while steering them in a more constructive direction.
China is a wildcard. China finds itself caught in the middle as an ally of Russia that also has friendly relations with Ukraine and deep economic ties to the countries supporting Ukraine. Currently, China appears to be weighing whether or how to support Russia in its war. An insightful commentary on this can be found here. How China proceeds is an important dimension of this to follow.
Finally, the Ukraine war provides some general insight into the politics of global catastrophic risk and wider sets of policy issues. The war has been an extreme example of the “policy window” phenomenon, in which sudden events create the possibility of massive policy change that would not have otherwise been possible. Dramatic shifts in European energy policy are an example. In contrast, policy windows can be more elusive for issues that move more gradually, such as many issues involving environmental or technological change. Whenever policy windows open, they can be acted upon most successfully by pre-existing communities of policy advocates. This speaks to the value of supporting such communities on an ongoing basis so that they are on standby for when policy windows open, especially communities that are interdisciplinary enough to be able to adapt their causes for a range of different policy windows.
One motivation for one of the probability forecasting exercises was to inform personal decisions of whether to evacuate major cities due to the risk of nuclear war. I would tend to de-emphasize probability estimates for these decisions due to the tenuous nature of the estimates (as outlined above) and due to the range of other important information.
First, anyone whose evacuation costs are very low might as well do it. This could apply, for example, to people with a vacation home in a remote area that they would enjoy spending some more time at, or to living in cities that may be worth evacuating for other reasons, such as cities in Ukraine or even Russia. The odds of a nuclear attack on a major city may be low, but for some people it might make sense to relocate just in case. The current risk of attacks on major cities seems driven primarily by a nuclear attack resulting from a massive failure of early warning systems resulting in inadvertent nuclear war. That could occur at any time and is more likely now given the high geopolitical tensions.
Second, for everyone else, I would recommend making some basic preparations and plans and monitoring for changes. These are low-cost activities that arguably one should be doing anyway. Preparations could be the same basic emergency preparations that are applicable across a wide range of catastrophe/disaster scenarios, including local natural hazards; see e.g. ready.gov. Plans would cover what to do in the event that nuclear war appeared imminent; this could include planning to make more detailed plans later. Monitoring could include keeping an eye on the news or following social media from experts such as Hans Kristensen and Pavel Podvig. The two of them are specifically experts in the composition of nuclear forces. If any changes are occurring indicating that nuclear war may be getting closer, they would be among the first to report it publicly. In a recent interview, Kristensen said “There is not a significant risk at this point” because “there are not significant changes on the ground for really doing that.”
Given the rapidly changing events, it may be difficult for philanthropy to constructively intervene to reduce the immediate nuclear war risk—not necessarily impossible, just difficult. Interested philanthropists are welcome to contact me about specific opportunities.
More generally, the current situation demonstrates the value of maintaining a robust field of nuclear security on an ongoing basis. I have been very impressed by the quality of discussion produced by the nuclear security community regarding the war. Indeed, on a number of occasions I have abstained from speaking out because they were already making the same points better than I would have been able to. With the MacArthur Foundation winding down its nuclear security program, I worry that the nuclear security field will be less able to effectively respond to the next crisis. MacArthur’s departure creates a void that I believe would be worth filling, and I am glad to see some funders taking this on.
In addition to general support for work on nuclear security, I see particular value in funding two types of work at the intersection of nuclear weapons and global catastrophic risk. First, basic research on nuclear war risk can clarify how much nuclear war should be prioritized within the overall portfolio of work on global catastrophic risk and what specific nuclear war interventions to prioritize. Second, support for people with global catastrophic risk backgrounds could help them get more involved in nuclear security, especially to achieve cross-fertilization between the fields of global catastrophic risk and nuclear security. This can include learning lessons from nuclear security for other global catastrophic risks such as artificial intelligence. Therefore, cross-fertilization has value for addressing both nuclear war risk and other global catastrophic risks. Nuclear war risk may be quite large, but even if the risk turns out to be low, it can still be worthwhile to support work at the intersection of nuclear security and global catastrophic risk.
My Own Activities
Thus far, I’ve taken a few specific actions regarding the Ukraine war.
To reduce the immediate nuclear war risk, I’m focusing on activities that draw on my own particular expertise while avoiding the numerous activities that other people are already doing at least as well as I could. First, I’ve done some outreach to reduce the risk of inadvertent nuclear war due to false alarms from asteroids and space weather. It’s an obscure risk that I happen to have background on; see this and this on asteroids and this on space weather. Because it’s more obscure, fewer people are focused on it. File it under “low importance, but high tractability (for me) and high neglectedness”. Through correspondence on asteroid false alarms, I’m now under the impression that this risk is even lower than I previously thought it was. I hope to be able to share details of this publicly at some point.
Second, I’m doing some work to bring a long-term, global catastrophic risk perspective to bear on the dilemma faced by NATO of how to balance aggressive support for Ukraine against the risk of escalating the conflict. This is arguably the single most important policy decision for NATO regarding the ongoing war. Details of this are also not yet publicly available.
Other topics that currently have my attention include scenarios for how the war will end; China’s role; postwar reconstruction and accompanying narratives, especially Russian narratives of humiliation; the role of epistemic communities in nuclear crises; various policy windows, including unconventional ones such as on tax havens; food prices and their implications for political stability, especially in poorer countries; and lessons for future nuclear war risk and other global catastrophic risks.
For anyone pursuing a deeper dive, here are some good readings across a range of relevant topics. I don’t agree with everything being said, but they’re nonetheless informative.
Nuclear war: general discussion: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
Nuclear war: Russia attacking Ukraine: 1, 2, 3, 4.
Nuclear war: probability estimates: 1, 2, 3, 4, 5, 6.
War (conventional or nuclear) between NATO and Russia: 1, 2, 3, 4, 5.
Russian nuclear force posture: 1, 2, 3, 4.
Brinkmanship & escalation: 1, 2, 3, 4, 5.
Endgames for the war: 1, 2, 3, 4.
Russian regime stability: 1, 2, 3, 4, 5, 6, 7, 8.
Future international cooperation: 1, 2.
Russian humiliation narratives: 1, 2, 3.
Thank you, Seth, for a great piece. I have devoted most of my 30year-odd career to nuclear diplomacy, deterrence and acquisition policies. What I have consciously steered clear of up until now has been scenarios for nuclear use, de-escalation strategies, forecasting, consequences and resilience. This whole experience has had me wondering whether my focus at CSER would benefit from that focus, and what that would look like in research, analysis and efforts to impact. I'm thinking that in the short term we need to look at escalation scenarios, de-escalation options at each stage (off-ramps) for both sides, as well as the consequences of limited or less than limited nuclear exchange.
If you're looking for co-conspirators on your second line of work, Seth, I'd be very keen to talk further. I do think this situation has to be seen as an opportunity, a point of inflexion, to bring attention to some of the dangerous postures all nuclear weapon states engage in.
Paul Ingram, CSER
This a really useful set of thoughts and further readings, thanks for sharing Seth. I especially liked your points on policy windows, and on cross-fertilisation between the nuclear and GCR fields.
Shared the Twitter thread below on Feb 28, crossposting as similar reflections:
A few thoughts on global catastrophic risk a few days into Putin's invasion - would be very interested in other's reflections!
I found this valuable, thank you.