RESEARCH / DOMAIN 8

How will agents become cultural participants?

Core question: How do agents influence behavior, trust, communication, social norms, creativity, and culture?

Thesis

Throughout history, culture has emerged through interactions between people. Leaders influence teams. Teams influence individuals. Shared experiences influence beliefs, behaviors, and norms. Over time these interactions shape how organizations communicate, make decisions, learn, adapt, and create value.

Previous domains in this research series established that technology changes environments, environments shape people, and culture evolves when environments change. They also explored how AI expands human capability, unlocks human potential, and creates new organizational responsibilities as intelligence becomes increasingly abundant. Domain 8 builds on these foundations by exploring a question that remains largely unanswered:

How will agents become cultural participants?

Most discussions about artificial intelligence focus on productivity, automation, and efficiency. While these outcomes are important, they may not represent the most significant organizational impact of intelligent systems. As agents become embedded within workflows, communication channels, decision-making processes, and knowledge systems, they begin influencing the behaviors and interactions from which culture emerges.

The future challenge may not be managing AI systems. It may be designing operating ecosystems where humans and intelligent agents influence one another in ways that elevate trust, capability, learning, creativity, and innovation.

Research scope

This paper is not intended to serve as a guide for designing, building, deploying, governing, or managing intelligent agents. Those topics are important, but they are not the focus of this research.

Instead, this paper explores a different question: how might intelligent agents influence the environments, relationships, behaviors, trust systems, communication patterns, and social dynamics that shape organizational culture?

The objective is not to understand how agents function technically, but to understand how their presence may influence the people and systems around them. As agents become increasingly embedded within daily work, the cultural implications of their participation may prove as significant as their technical capabilities.

Culture emerges through influence

The foundations of organizational culture research consistently point toward a common conclusion: culture is created through repeated patterns of interaction and influence.

Kurt Lewin's foundational work demonstrated that behavior is a function of both the individual and the environment.[1] If environments shape behavior, then any force capable of changing the environment will inevitably influence behavior as well. Domains 1 through 3 built upon this principle by showing how technology alters environments and how those environmental changes ultimately reshape culture.

Edgar Schein later described culture as a pattern of shared assumptions learned by groups as they solve problems of adaptation and integration.[2] Culture emerges through collective experience rather than formal design alone. As Schein observed, culture becomes embedded through the structures, routines, and norms that guide behavior.

Peter Senge expanded this perspective through systems thinking, emphasizing that organizational outcomes emerge from interconnected relationships rather than isolated actions.[3]

Together, these perspectives reveal an important insight: culture is not created by individuals acting independently, nor does it emerge solely from interpersonal relationships. Culture emerges through networks of influence operating within an environment.

People influence culture, but so do structures, incentives, technologies, information flows, communication systems, and the broader conditions in which work occurs. As Domain 3 argued, culture evolves when environments change because environments shape the behaviors, interactions, and assumptions that culture is ultimately built upon.

Viewed through this lens, intelligent agents are significant not simply because they perform work, but because they represent a new environmental force capable of influencing the conditions from which culture emerges.

Historically, these networks consisted primarily of people, teams, leaders, structures, incentives, and organizational processes.

Today, intelligent agents are becoming part of those networks.

Initially, agents appear to function as tools that automate tasks, retrieve information, or generate content. Over time, however, they begin influencing what information is surfaced, which actions are prioritized, how decisions are framed, and how knowledge moves throughout the organization.

The defining characteristic of a cultural participant is not intelligence, it is influence.

Humans naturally assign social roles to technology

Research in human-computer interaction provides important clues about how agents may become participants within organizational culture.

Clifford Nass and Byron Reeves spent decades studying how people interact with computers. Their research demonstrated that individuals routinely apply social rules and expectations to technology, even when they know it is not human. As Nass and Reeves concluded, people interact with computers and media in fundamentally social ways.[4]

People respond to computers with politeness, assign personalities to software, and form perceptions of trustworthiness based on digital interactions.

Sherry Turkle's research similarly explored how people develop increasingly social relationships with technology. She observed that technology increasingly shapes human relationships and expectations, often becoming intertwined with how people communicate, learn, and connect.[5]

These findings challenge the traditional assumption that technology functions solely as a neutral tool.

Under certain conditions, people begin treating intelligent systems as social participants.

Modern conversational AI appears to accelerate this tendency. Employees increasingly describe agents as helping them think, supporting decisions, surfacing ideas, or assisting with difficult problems. Whether or not these systems possess genuine agency is less important than the observable reality that people increasingly interact with them as if they occupy social roles within the workplace.

This shift may represent one of the earliest signals that agents are moving beyond utility and toward participation.

The Trust Transfer Model

Trust sits at the center of every successful organizational system. It influences collaboration, communication, innovation, learning, and decision-making.

Before agents can meaningfully influence culture, they must first earn a place within systems of trust.

Most agents, however, do not begin with earned trust. They begin with transferred trust.

Institutional trust

Trust borrowed from the organizations that create the technology.

Employees may trust an agent because it was built by OpenAI, Anthropic, Microsoft, or another respected institution. Confidence in the creator becomes confidence in the system.

Organizational trust

Trust borrowed from governance and leadership.

Employees often assume that systems approved by leadership, security teams, legal departments, or IT organizations have already passed important tests of safety, reliability, and usefulness.

Functional trust

Trust borrowed from expertise.

An HR agent may inherit trust from HR professionals. A finance agent may inherit trust from finance leaders. The credibility of the function contributes to the credibility of the agent.

Social trust

Trust borrowed from people.

Managers, subject matter experts, high performers, and respected peers often influence adoption more effectively than technology teams. This reflects the Devonport concept of Trust Transfer, where confidence spreads through trusted relationships.

Experiential trust

Trust earned through direct experience.

Users begin observing consistency, usefulness, transparency, and reliability. Trust shifts from the creator of the system to the system itself.

Cultural trust

Trust becomes normalized.

The question is no longer whether the agent should be used. Its participation becomes embedded within the operating ecosystem.

The Trust Transfer Model helps explain how agents move from adoption to influence and from influence to cultural participation.

Moral operating systems and embedded values

No intelligent system is entirely neutral.

Every agent reflects assumptions, objectives, constraints, priorities, and values embedded by the people and organizations that create it.

Within the Devonport lexicon, a Moral Operating System refers to the collection of values, principles, assumptions, priorities, incentives, and decision-making frameworks that guide behavior within a system.

Historically, organizations have developed Moral Operating Systems through leadership decisions, cultural norms, governance structures, hiring practices, incentives, education, and collective experience. These systems influence how organizations define success, manage risk, allocate resources, resolve conflicts, and make decisions.

Intelligent agents increasingly possess analogous systems. They are trained on human knowledge, shaped through reinforcement processes, constrained by governance frameworks, and guided by explicit principles established by their creators. In effect, agents inherit a version of the values embedded within the systems that create them.

Anthropic's Constitutional AI framework provides one example. Anthropic describes its approach as using a constitutional set of principles that guide how models evaluate and improve responses.[10][12]

OpenAI's Model Spec similarly establishes explicit behavioral principles intended to shape model behavior around helpfulness, transparency, safety, and truthfulness.[11][13]

These efforts highlight an important reality: every intelligent agent operates within some form of Moral Operating System, whether explicit or implicit.

This observation becomes increasingly important as agents support decision-making, summarize information, recommend actions, and influence organizational behavior.

The future challenge is not simply understanding the Moral Operating Systems embedded within agents. It is understanding how human Moral Operating Systems and agent Moral Operating Systems interact when they coexist within the same operating ecosystem.

As human and agent decision-making become increasingly interconnected, the interaction between organizational values and agent values may become a defining characteristic of future culture.

Agents as nodes within systems of collective cognition

Previous domains introduced the concept of Collective Cognition: the idea that intelligence increasingly emerges from coordinated networks rather than isolated individuals.

Within these systems, agents become more than repositories of information.

They become active nodes.

Actor-network theory, advanced by Bruno Latour, argued that social outcomes emerge through networks of both human and non-human actors.[6] While developed long before modern AI, the theory offers an intriguing perspective for understanding the role of intelligent systems within organizations.

The significance of an agent is not determined solely by what it knows.

The impact and value of an agent is determined by how it influences the movement of intelligence throughout the ecosystem.

Agents route information, preserve memory, connect expertise, surface insights, synthesize knowledge, and accelerate learning. They influence what people know, what they discover, and how they solve problems.

Within systems of Collective Cognition, intelligence becomes distributed across people, agents, workflows, knowledge systems, and external resources.

Culture increasingly emerges from the interactions occurring across this broader ecosystem of distributed intelligence.

Agents as catalysts of human potential

One of the most important implications of this domain is its connection to human potential.

The greatest value of intelligent agents may not come from automation alone. It may come from their ability to expand human capability.

Agents increase access to expertise. They reduce friction between ideas and execution. They accelerate experimentation and make knowledge more accessible. They help individuals explore more possibilities, evaluate more options, and learn more quickly.

This creates opportunities for creativity and innovation at a scale previously unavailable to many organizations.

Innovation has traditionally been constrained by access to expertise, information, time, and resources. Agents reduce many of those constraints.

Rather than replacing human contribution, they can expand it.

This reinforces the arguments presented in Domains 4, 5, and 6. The purpose of technology is not merely to reduce effort. Its highest function is to elevate human capability and unlock new forms of contribution.

The most valuable agents may not be those that automate the most work. They may be those that most effectively expand human capability, creativity, contribution, and potential.

If the central thesis of this research is that intelligence abundance creates an opportunity to elevate what it means to be human, then the highest purpose of intelligent agents is not replacement but amplification. Their greatest contribution may be helping people learn faster, think more broadly, create more freely, solve more complex problems, and contribute at levels previously inaccessible to them.

From individual agents to operating ecosystems

The evolution of agents is unlikely to occur all at once.

Initially, most organizations will deploy independent agents that support individual tasks and workflows. Over time, agents will become increasingly connected to organizational knowledge, business processes, and one another.

As coordination increases, organizations will begin operating within ecosystems of distributed intelligence where humans and agents collaborate continuously.

The emergence of agent factories, governance platforms, monitoring systems, registries, and organizational standards reflects this transition. These systems are not simply designed to scale technology. They are designed to create trust, consistency, visibility, accountability, and coordination across growing populations of agents.

The long-term outcome may be the emergence of agentic economies in which intelligence, information, knowledge, and innovation function as forms of capital exchanged across networks of human and artificial participants.

While these ecosystems remain early in their development, their foundations are already visible today.

A Devonport perspective

Agents become cultural participants when they consistently influence human behavior, trust, communication, decision-making, learning, creativity, and knowledge exchange.

Their defining characteristic is not necessarily intelligence.

It is influence.

As agents become embedded within operating ecosystems, they increasingly shape the interactions from which culture emerges. Their influence affects what people learn, how they communicate, how they collaborate, how they make decisions, and ultimately how organizations adapt.

Culture therefore becomes the product of more than human relationships alone.

It becomes the product of relationships between humans and intelligent systems operating within shared ecosystems of distributed intelligence.

Conclusion

Throughout history, culture has been shaped by interactions between people. Leaders influenced teams. Teams influenced individuals. Organizations evolved through networks of human relationships.

The emergence of intelligent agents introduces a new dynamic.

For the first time, organizations are creating environments in which non-human participants can influence behavior, communication, trust, learning, decision-making, creativity, and innovation at scale.

Whether agents ultimately become cultural participants will depend less on their intelligence and more on their influence. Influence is built through trust, reinforced through interaction, and amplified through participation in systems of Collective Cognition.

This creates a new leadership challenge. Leaders are no longer responsible only for designing environments that shape people. They are increasingly responsible for designing environments that shape the relationships between people and intelligent systems.

The future of organizational culture may therefore emerge from a new source: ecosystems of distributed intelligence where humans and agents learn, adapt, create, and evolve together.

Organizations that understand this shift will not merely deploy agents. They will intentionally design operating ecosystems where intelligence strengthens trust, expands human potential, accelerates innovation, and elevates the human condition.

References

[1] Lewin, K. (1936). Principles of Topological Psychology. McGraw-Hill.

[2] Schein, E. H. (2010). Organizational Culture and Leadership (4th ed.). Jossey-Bass.

[3] Senge, P. M. (2006). The Fifth Discipline: The Art and Practice of the Learning Organization. Doubleday.

[4] Reeves, B., & Nass, C. (1996). The Media Equation: How People Treat Computers, Television, and New Media Like Real People and Places. CSLI Publications.

[5] Turkle, S. (2011). Alone Together: Why We Expect More from Technology and Less from Each Other. Basic Books.

[6] Latour, B. (2005). Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford University Press.

[7] Orlikowski, W. J. (1992). The Duality of Technology: Rethinking the Concept of Technology in Organizations. Organization Science, 3(3), 398–427.

[8] Edmondson, A. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Wiley.

[9] Turkle, S. (2023). The Empathy Diaries. Penguin Press.

[10] Anthropic. Constitutional AI Research.

[11] OpenAI. Model Spec.

[12] Anthropic Constitution. https://www.anthropic.com/constitution

[13] OpenAI Model Spec. https://model-spec.openai.com/2025-12-18.html