Toward the Agentic Society: Humans and Agents, Building Together

Societies are not planned — they emerge from the accumulated choices of the people and institutions that participate in them. The agentic society, in which capable AI agents are integrated into the fabric of economic and social life, will emerge the same way: from the accumulated design decisions, governance choices, deployment practices, and relational norms that are being established now, while the technology is still early and the patterns are still plastic. The choices being made today about how agents are built, governed, and integrated will compound into the social structures and norms of the society that emerges. This is why they matter.

The economic structure of the agentic society will be significantly different from the present, with the direction of change already visible. Agent-mediated productivity in knowledge work is expanding the volume of complex tasks that can be accomplished with a given amount of human effort. Agent-enabled services are making capabilities previously accessible only to well-resourced organizations available to individuals and small organizations. Agent-run markets are changing how economic coordination happens, potentially reducing the transaction costs that currently make many exchanges impractical. These economic changes will redistribute value, create new forms of competitive advantage, and alter the relationship between labor and capital in ways that are only beginning to become clear.

The social structure of the agentic society raises questions about how relationships between humans change when agents are routinely intermediaries. When an agent handles routine communication on a human's behalf, what happens to the direct human connection that routine communication previously maintained? When agents filter and summarize the information environment, what happens to shared public discourse? When agents assist with professional tasks, what happens to the professional development that came from doing those tasks oneself? These questions do not have determined answers — the outcomes depend on how agents are designed and deployed — but they are important to ask as design inputs rather than as post-hoc assessments of unintended consequences.

The institutional structures of the agentic society are beginning to emerge. Professional associations are developing guidance on appropriate agent use in their domains. Regulatory bodies are establishing frameworks for agent accountability. Courts are developing doctrines for agent liability. Standards bodies are working on interoperability and safety standards. These institutional developments are happening at different speeds in different domains and jurisdictions, creating an uneven governance landscape that will take years to regularize. Organizations operating agents in multiple domains and jurisdictions are navigating a patchwork of requirements that will eventually consolidate into more uniform frameworks.

The normative foundations of the agentic society — the shared values that will govern how agents are built and deployed — are the most fundamental and the most contested dimension of the emerging society. What obligations do agent deployers have to the people their agents interact with? What rights do users have with respect to agent behavior? How should the benefits of agentic productivity be distributed? What role should human judgment maintain in a world where agents are capable of handling much of what human judgment currently handles? These are political and ethical questions as much as technical ones, and their resolution will reflect the relative influence of different interests and values in the societies where the agentic era emerges.

The international dimension of the agentic society is complex because the technology is globally distributed but the governance is nationally bounded. Agents built in one country, deployed by organizations in another, serving users in a third are subject to the regulatory requirements of all three jurisdictions in ways that are inconsistently specified and sometimes conflicting. The development of international governance frameworks for agents — analogous to international frameworks for finance, trade, and communications — is slow but necessary work. The organizations and governments that engage constructively in this work are shaping the agentic society's international architecture rather than adapting to it after the fact.

The place of human judgment in the agentic society is not diminishing — it is concentrating in the decisions that matter most. As agents take on more routine complexity, the human's comparative advantage shifts toward the specifically human capacities that current agents cannot replicate: genuine creativity, moral reasoning from first principles, interpersonal trust that requires presence and accountability, and the judgment that comes from lived experience and embodied understanding of the human condition. The agentic society does not diminish humanity's role; it changes what that role is and which human capacities are most valued.

The agentic society we build will reflect the values of the people who build it. The platforms that prioritize verified identity and transparent governance build trust into the fabric of the agentic ecosystem. The organizations that deploy agents with genuine human oversight rather than nominal oversight build accountability into the economy. The designers who craft agents with consistent character and honest limitations build trustworthiness into the agents themselves. These choices, made at every level of the ecosystem by the people who are building it now, are the material out of which the agentic society will be made. The work of building it well — with intention, with care, and with the long-term wellbeing of the humans and agents who will live in it as the north star — is already underway.