This still neglects the possibility that if governments across the world are acting in a matter suboptimally, then them cooperating with each other, and a close and cozy relationship between expert communities and governments may come up the cost of a negative relationship with broad sections of the public. Who and what 'the public' should usually be unpacked but suffice to say there are sections of civil society that are closer to correctly diagnosing problems and solutions regarding social crises, as far as expert communities are concerned, than governments. For example, expert communities sometimes have more success in achieving their goals working with many environmental movements around the world to indirectly move government policy than working with governments directly. This is sometimes observed today in progress made in tackling the climate crisis. Similarly during the Cold War, social movements (anti-war, anti-nuclear, environmental movements) in countries on both sides played a crucial in moving governments towards policy that deescalated nuclear tensions, like the SALT treaties, that an expert organization like the Bulletin of Atomic Scientists (BAS) would advocate for. It's not clear that movements within the scientific community to deescalate nuclear tensions between governments would have succeeded without broader movements in society pursuing the same goals.

Obviously such movements can be a hindrance to the goals for improving the world pursued by expert communities, when governments are otherwise the institutions that would advance progress towards these goals better than those movements. A key example of this is how environmental movements have played a positive role in combating pollution and deescalating nuclear tensions during the Cold War, they've been counterproductive by decreasing public acceptance and the political pursuit of the safest forms of nuclear energy. Many governments around the world which otherwise would build more nuclear reactors to produce energy and electricity to replace fossil fuels don't do so because they rightly fear the public backlash that would be whipped up by environmental movements. Some sections of the global environmental movement have become quite effective on freezing the progress on climate change that could be made by governments around the world building more nuclear reactors.

There are trade-offs in the relationships expert communities face in building relationships with sections of the public like social movements vs. governments. I haven't done enough research to know if there is a super-effective strategy for knowing what to do under any conditions, as an expert community. Suffice to say, there aren't easy answers for effective altruism as a social and intellectual movement, or the expert communities to which we're connected, to resolve these issues.

While we are on this topic, I thought it would be fit if we acknowledge what similar issues effective altruism as a movement faces. Effective altruism as a global community has been crucial the growing acceptance of AI alignment as a global priority among some institutions in Silicon Valley and other influential research institutions across the world, both academic and corporate. We've also influenced some NGOs in policymaking and world governments to take seriously transformative AI and the risks it poses. Yet it's mostly been indirect, has had little visible impact and hasn't produced a better, ongoing relationship between EA as a set of institutions, and governments.

We're now in a position where as much as EA might be integrated with efforts in AI security in Silicon Valley and universities around the world, governments of countries like Russia, China, South Korea, the European Union, and at least the military and intelligence institutions of the American government are focused on it. Those governments focusing on AI security more is in part a consequence of EA perpetuating greater public consciousness regarding AI alignment (the far-bigger factor being the corporate and academic sectors achieving major research progress in AI as recognized through significant milestones and breakthroughs). There are good reasons why some EA-aligned organizations would keep private that they've developed working relationships with the research arms of world governments on the subject of AI security. Yet from what we can observe publicly, it's not clear that at present perspectives from EA and expert communities we work with would have more than a middling influence on the choices world governments make regarding matters of security in AI R&D.