Comments
Introduction
The Manhattan Project's history is a controversial one. A group of elite scientists enable unprecedented destruction with the justification of opposing the existential threat of authoritarian lock-in, all while wrestling with the apocalyptic risk of atmospheric combustion. It's a compelling story, but the narrative often abruptly stops at 1945 with the bombing of Hiroshima and Nagasaki. You may get some snapshots of regretful physicists years after this date but the unified political struggle that continues directly after Imperial Japan's surrender is seldom mentioned. I propose that the early political history of nuclear weapons is in desperate need of revisiting as it provides a unique look into the possibilities for global reordering that follow dramatic technological change and the 'spectacular moments' that follow. Learning from the movement for world federalism supported by many nuclear scientists in the mid-1940s, considering both its successes and failures, will provide a useful reference for how to push for positive international change during the advent of AGI. Perhaps, with this case study as a precedent, it will even be possible to steer humanity into a better future without mass destruction sparking global debate.
Whistlestop Political History of the Nuclear Age
Serious consideration concerning the politics of the post-nuclear world order began early, before even the success of the Trinity test. A loose coalition of physicists saw the necessity for world government control over nuclear resources, and therefore weaponry, if peace were to be sustained after the defeat of the Axis. In the literature on the topic the seed of this idea is traced back to a group of scientists and engineers in Oak Ridge, Tennessee who began a "Letters on World Government" project which would see them write to hundreds of Anglo-American leaders in science and government spreading their proposal. However, while considered, it wasn't until the spectacular moment of the Nagasaki and Hiroshima bombings that the question entered the public conscience and the world reckoned with the issue that scientists had been quietly considering for months to years.
Tentative planning on topics of public education on nuclear issues and international governance as expressed in documents such as the 1944 "Jefferies Report" authored by the likes of Enrico Fermi and other Manhattan Project physicists were suddenly actioned. Public intellectuals began to express opinions in line with the Manhattan physicists with "The Bomb and Civilization" being published by Bertrand Russel on the 18th of August 1945. He would later go on to work with Einstein, publishing joint manifestoes and founding the Pugwash Conferences, with the aim of inspiring new political thinking to ensure humanity's survival.
Initially, the success of this movement was immense. Despite nuclear weaponry naturally falling into military control at first, the established powers had no idea as to how the former world order could be maintained with such revolutionary technology. President Truman commissioned a report on new approaches to governance by his Secretary of State Dean Acheson (after Henry Stimson's resignation from the project), David Lilienthal, and luminaries from the newly established Federation of Atomic Scientists with the main representative being Oppenheimer. The organisation was staunchly in favour of global control of nuclear resources, a sentiment which was communicated in the final report. It proposed all fissile material be controlled by an international "Atomic Development Authority". At first, post-war collaboration of this kind was surprisingly warmly received. The planned United Nations Atomic Energy Commission was supported by both the USA and USSR which would be a precursor to international control of nuclear energy. The further steps put forward in the initial draft of the "Acheson-Lilienthal" report would complete this process.
However, the final proposal, informed by the aforementioned report, written by Truman's representative to the now-established UN Atomic Energy Commission Bernard Baruch would begin to encounter difficulty. The paper that would be presented before the UNAEC in 1946 reduced the scope of global control that the initial draft claimed as necessary, also including several amendments from US Congress made in nationalist self-interest. Baruch was infamous among the Federation of Atomic Scientists for his reluctance to listen to their concerns as well as obstinance in his ideas. Disagreement between Western powers and the USSR over pre-emptive American nuclear disarmament as well as inspections by what the Soviet Union saw as a biased United Nations sunk the initial plan. A truly multinational authority was seen as a pre-requisite for a deal and as tensions rose between the two superpowers the appetite for one waned. There would be changes to the report and Soviet counter-proposals over the next year but none would see success.
In the meantime, the head-start that wartime secrecy had afforded the nuclear scientists in their planning and advocacy had been spent and dissenting voices in the political-strategic world began to form a coherent alternative. It began as simple attacks on world government itself and the feasibility of US-Soviet cooperation for practical and nationalist reasons, meant as a direct challenge to Einstein's publicity tour spreading his globalist ideas after the war. However, before long these critiques would coalesce into what we now call deterrence theory. The first work in this school is credited to Bernard Brodie's “The Absolute Weapon: Atomic Power and World Order” which explored ways of maintaining independent, sovereign statehood with widespread nuclear proliferation. The article also put forward the idea of a nuclear peace as a natural consequence of this fact. Between rising tensions, the path dependency of the pre-existing world order, and classic coordination failure the idea of something dramatic and novel being essential for humanity's survival dwindled. What was understood as necessary only a year or two earlier was now popularly perceived as unworkable or plain naive. Within the US the death knell of the world government movement can be said to have come with its attachment to former Vice President Henry Wallace who ran for office under the banner of the newly-formed Progressive Party. Controversy sunk his campaign and any hope of meaningful change with it. Many of the organisations formed toward the goal of international nuclear collaboration would soon disband or find themselves in a lull.
Lessons Learned
It would appear that the two greatest strengths of the advocate scientists were the foreknowledge of nuclear dangers and the head start in planning which this gave them, as well as the access to high-level decision-makers. The former could also be seen as a minor weakness as wartime secrecy impeded public advocacy, however the limited benefits that press tours and public statements afforded the movement make this seem relatively minor. If we are considering awareness-raising then the spectacular moment of Hiroshima and Nagasaki is unparalleled in stirring debate and drawing attention to the need for change. On the other hand, the movement for world nuclear government was clearly impeded by a lack of a clear policy proposal to coincide with this moment. This delay invited state governments into the process of drafting policy from scratch which opens up the task to nationalist roadblocks and bureaucratic obtuseness. With every passing day the chronological and mental distance from the spectacular events grows and political rivals have opportunities to rally in defense of the old order. This loss of momentum can be observed in unsuccessful political movements globally, causing conservatism to rule out by rejecting new ideas or co-opting them.
What does this have to do with AI?
Artificial Intelligence technology is radically transformative. I doubt that I need to defend that premise in this space. However I would argue, if AI can scale to AGI or even superintelligence, that we haven't yet had a spectacular moment in the sense that the 1945 atomic bombings were for nuclear technology. In many ways this is related to the idea of a "warning shot"; a near miss or show of destructive potential which raises questions about the safety of a technology. There are plenty of good arguments as to why we may never get the chance to be warned of AI's capabilities in a way that's politically meaningful, either through flawed human perception or a deadly intelligence explosion. I propose that the idea of a spectacular moment can be distinct from this. Although a warning shot is naturally spectacular, there is the possibility of creating a moment of widespread public debate without a hostile deployment of deadly and volatile technology. This idea has some precedent in securitisation theory if any EA's wish to further explore how dramatic events, or the threat of one, can permit approval of sweeping executive powers.
So, when compared to the concerned scientists of the Manhattan Project, we can be said to currently be just after the Trinity test. Initial trials of Artificial Intelligence have seen dramatic successes. Currently the US has a monopoly on the most advanced forms of this technology but it is widely acknowledged that it won't be long before the capabilities demonstrated in these initial tests will be proliferated further with the potential for steep scaling. The concept of AGI is just beginning to enter the public sphere and with it the dream of widespread use for the good of humanity. Despite this, the destructive capabilities that accompany it are not yet known or widely acknowledged. At this stage in 1945 the idea of a different world order befitting the nuclear age was present, but only just forming. The bombings of Hiroshima and Nagasaki would accelerate this discussion and present an opportunity for meaningful change, but from my reading it still appears that the scientists of the day were caught off-guard by this moment.
My central point to this essay is that plans for how to govern a post-AGI world need to be made immediately. Whether it be centralised and monitored compute or the restriction of certain technologies/models, a clear proposal needs to exist so as to not replicate the failures of 1945. Not vague suggestions but concrete policy. Ideas that may appear dramatic or utopian are encouraged if they ensure the survival of humanity and the avoidance of existential danger. The federal funding of AI projects and the intimate involvement of AI pioneers with world leaders can be compared without much compromise to the situation of nuclear physicists in the 1940s. Before competing world views had the chance to join in resistance even the most optimistic of proposals were considered when delivered by these field experts.
Of course, as mentioned previously, this was partly as a result of the spectacular moment. It is of utmost importance that a disaster for human dignity on the scale of the 1945 atomic bombings is avoided. This is the second goal of comparing our present situation to that of the Manhattan Project. Ideally, we can receive similar results and attention through mass publicity or a controlled demonstration of the technology we are developing's destructive potential, creating a spectacular moment without the consequences. However, if history regrettably repeats itself then we must be ready to propose dramatic and meaningful change before the world can settle into its old patterns.
We have several advantages when compared to the scientists of 1945. Namely, better access to the public through social media, experience with global cooperation in the face of existential threats, as well as the internet which provides us with the possibility to research and collaborate at a greater scale and pace than could have been imagined 80 years ago. With the benefit of hindsight we can, and must, do better.
Personally, I am not in a position to kickstart this kind of movement or begin a project of this scale. There's a chance that this strategy of change isn't truly robust for several reasons, be it related to authoritarian lock-in or unexpected technological factors, but my research tells me that this is the best chance of avoiding some rather nasty futures. Regardless, it is pretty objective that there are meaningful lessons to be learned from the experiences of the early nuclear era and even if you come to different conclusions than myself I strongly support more attention being paid to this period in history. More granularity on matters such as the effectiveness of publicity tours, modes of communication with federal government and internationally, as well as organisation structure of nuclear advocacy organisations are needed if 1945 were to be used as an effective model.
TLDR - Parallels between Manhattan Project and current AI development make it a useful case study for investigating how radical technological change can alter the global order. These moments caused as a consequence of this technology often provide windows of opportunity for dramatic positive change. We should use the political history of the early atomic age as a case study for how to effectively navigate these periods and avoid the associated existential risk.
Sources Used
