The epistemic community research programme, which in the last generation helped frame International Relations understandings of the relationship between knowledge and power, rested on the influence of scientists and experts on policymaking through framing, persuasion, and socialization. In this chapter, we argue that while pioneering, the epistemic-community research programme was incomplete, since it neglected the relationship between knowledge, power, and practice. Because practices are at the core of what epistemic communities are, do, and aim to achieve, we suggest a pragmatist practice-based approach, according to which knowing requires active participation in social communities and knowledge is not a product but is bound with action. We, therefore, explain the political adoption of knowledge-generated practices by the very nature of practicing and joining communities of practice. With this purpose in mind, we propose understanding epistemic communities as epistemic communities of practice. By identifying epistemic communities as a special and heuristically important case of communities of practice, we will open new and exciting avenues of theory-making and empirical research. We illustrate our approach by examining the establishment of nuclear arms control verification practices during the Cold War and the recent spread of a populist ‘post-truth’ epistemic community of practice.

This conclusion is but a new beginning. I start off engaging with the contributors’ friendly arguments and criticisms, many of which I adopt for future research. Looking forward, I demonstrate the practical potential of a world ordering research program, arguing that tackling world order’s urgent global vulnerability threats, such as climate change and Artificial Intelligence (AI) cannot wait for dependable global governance, let alone world government. Instead, I suggest cognitive-evolution politics. The latter involves, first, grasping that world order vulnerability results from the performance of simultaneous and synergetic local, national, and international, harmful practices. Second, cognitive evolution politics requires mutual trust and adopting global cooperation by necessity, which, in turn, calls for embedding global problems and solutions in national interests. Cognitive evolution politics may promote “embedded globalism” by (a) learning, contestation, experimentation, and negotiation within and across communities of practice; (b) power, understood as a decentralized and relational field of forces distributed across communities of practice’s multiple--state, corporate, scientific, and public opinion--stakeholders, and (c) practical democracy, consisting of democratic practices that sustain common humanity values while promoting national interests. I briefly illustrate my claims focusing on AI and communities of practice’s efforts to mitigate its harmful global consequences.