By learning from philanthropy’s history of building coalitions and advancing global coordination, John Palfrey, President, reimagines the AI arms race from competition to collective responsibility and outcomes that serve the common good.
Philanthropy at its best begins with a steadfast commitment to human dignity and to the long arc of the public good. It operates on the understanding that meaningful change rarely comes from rapid, reactive, short-term-oriented measures. Instead, it values prudence over speed, collaboration over competition, and the deliberate work of building systems that are just, resilient, and capable of sustaining positive outcomes over time. In this way, philanthropy is most powerful when it is oriented toward enduring impact: a focus on the health of people, institutions, communities, and societies as a whole.
These values are antithetical to the rhetoric of any kind of arms race, which centers velocity, dominance, and near-term advantage. The MacArthur Foundation’s mission statement is to build a more just, verdant, and peaceful world, yet today’s AI arms race narrative encourages a mindset of scarcity and competition, where the question is not how AI can serve humanity, but who can win first. A race framed around escalation rather than responsibility risks deepening inequities, accelerating environmental harm and climate change, and undermining collective security. We are already seeing how AI systems can disrupt livelihoods and concentrate power, and an unchecked contest for dominance can—and does—magnify those harms.
Philanthropy is uniquely positioned to model a shift in the focus from technological dominance to collective human security and equity.
The current policy debate is often seen as a binary choice between supporting AI innovation and competitiveness, or prioritizing other interests like security, jobs, and equity. We should be able to achieve at least a proportion of all of these goals for everyone. But if we follow our current trajectory, we risk excluding the interests of an extraordinary number of people, with just a small number of countries using AI to drive a massive consolidation of authority. By viewing the prevailing story as fundamentally incompatible with our mission, philanthropy is uniquely positioned to model a shift in the focus from technological dominance to collective human security and equity.
Collaboration at the Center
A foundation could exhaust all its resources in a year and still fall short of meeting the scale and complexity of this moment. So we do almost none of this work in isolation. If an idea is exciting enough for us to pursue, it usually means we need partners: others with the expertise, lived experiences, and perspectives we lack. No single organization holds all the answers, and working together allows philanthropy to scale both vision and impact. By privileging collaboration over competition, we confront AI’s complexity in ways that are more effective, just, and enduring. This coalition-driven approach stands in fundamental opposition to an arms race mentality that prizes unilateral speed and dominance over shared responsibility and inclusive impact.
When MacArthur considers whether to enter a new domain of work, we start with a simple but essential set of questions. Is this a meaningful problem in the world? Is philanthropy well-suited to play a constructive role? The challenges of AI, ranging from ethical and social implications to rapid technological change and global coordination, make these questions especially urgent and complex—and these are precisely the kind of challenges no single funder can address alone. Rather than chasing unilateral dominance, we build partnerships, invest in long-term solutions, and center human welfare to address complex, global challenges like AI. Whereas we are currently seeing nation states insist that issues of strategic importance must be handled alone, foundations know that our impact grows when we work together. We know that collaboration is not optional, but the only way to meet challenges of this scale; our model should be a blueprint for global powers, rather than a counterpart.
Collaboration is not optional, but the only way to meet challenges of this scale; our model should be a blueprint for global powers, rather than a counterpart.
Even as a proud American who believes in the manifold strengths and continued promise of the United States, I must acknowledge that our track record on protecting marginalized communities and children from the effects of emerging technology is uneven at best, and we should not assume that US legal norms are sufficient to guide global AI governance alone. The implications of this technology are worldwide, and our response must be as well. Foundations must make a conscious effort to support work beyond our borders and to invest in global collaborations.
Although we cannot meet every need, we can choose areas where our resources and independence enable us to support ideas and communities that might otherwise be overlooked – which, in turn, will serve the common good.
Applied to AI, this long-term orientation reinforces a core lesson from philanthropy: when we center people rather than power, we make better choices. By investing in humanistic, community-rooted approaches to AI, we can help ensure that the design and governance of these systems reflect democratic values rather than the logic of a technological race.
Learning from Philanthropy’s Past
Looking through the lens of history, we see that even the most complex technologies have been guided, and sometimes rescued, by coordinated action, with philanthropy playing a pivotal role in injecting the capital that helps shape global norms and policy. It is instructive to step back and consider the historical parallels to the challenges we face with AI. Philanthropy has intervened in these moments: funding research, convening international bodies, sustaining long-term coordination, and bridging gaps where political or institutional incentives fall short. It is a model for how these global challenges, no matter how complex, can be addressed through shared vision and effort.
The massive collaboration embedded within climate work is a striking illustration of what coordinated action—and national leadership—can achieve. The Intergovernmental Panel on Climate Change, for example, was born from a coalition of philanthropic and state support, demonstrating how shared resources and vision can create enduring global institutions. Similarly, the hundreds of millions of dollars in American philanthropic capital that funded the Global Methane Hub came directly out of COP26, the UN-hosted international climate summit and a key moment of global cooperation. It drew commitments from all the major emitting countries, with foundations stepping in to support countries that did not believe they had the technical capacity to tackle these challenges alone. American leadership in this case generated tangible climate benefits for the entire planet, showing how targeted action from one actor can catalyze collective impact and shared progress. Climate change is a defining issue of the future, affecting the entire planet, though some regions and populations will feel its impacts more acutely than others. The same is true of AI.
We have seen significant strides in cooperative, multilateral efforts to govern AI responsibly.
Yet at the heart of the AI arms race narrative is the premise of AI as an unprecedented technology, to which the principles of cooperation and multilateralism that have bridged global issues in the past are inapplicable. From the public narrative, one might imagine that global collaboration on AI is simply not possible and never has been. We know that this isn’t true of AI coordination in the past or the present. In reality, we have seen significant strides in cooperative, multilateral efforts to govern AI responsibly. The first intergovernmental standards on AI were adopted by the G20 in 2019, establishing a shared baseline for principles such as transparency, human oversight, and accountability. A year later, France and Canada spearheaded the Global Partnership on AI, bringing together governments, civil society, and industry from around the world to promote the responsible development and deployment of AI. At the Paris AI Action Summit in February 2025, philanthropies, including the MacArthur Foundation, joined governments and tech firms to launch Current AI, a $400 million public‑interest foundation focused on global collaboration on AI Philanthropic funding, including support from MacArthur, helped launch Humanity AI later in the year, linking the AI ecosystem across government, academia and civil society to support long-term coordination across sectors towards a people-driven future.
In recent years, it is the philanthropic sector that has stepped in to model how AI can be aligned with human rights and democratic values, showing that deliberate, thoughtful investment can guide a rapidly evolving technology toward outcomes that serve the broader public good. But it doesn’t have to act alone.
Humanity Above All
The word philanthropy derives from the Ancient Greek: “love of humanity.” The broader sweep of humanity must have the opportunity, and the power, to shape the future. The path to meaningful impact and to being proud of our contribution lies in elevating better ideas, grounded in what people want from technology and from one another. This is a design moment—for the technologies, for the way we relate to one another across geographic and political boundaries, and for what the future looks like.
Our role is to help build the networks, institutions, and public voices that can assert these values, so that technological progress serves the common good rather than overwhelming it. The philanthropic model, and lessons from its past and present, can instruct us in an alternative approach in which we refuse the dominant narrative in favor of something better. A zero-sum race reimagined as a collective pursuit of shared progress that will guide us to a brighter future with AI as a catalyst, not a threat.
This piece first appeared in AIxGEO’s Reimagining the AI Arms Race, an anthology of essays.
