Can AI Help with Consensus-Making?

Updated: April 17, 2026

Key Points

• The costs of acquiring and exchanging information help determine the hierarchical structure of organizations.

• GenAI may improve decision speed, strategic alignment, and consensus, while also creating new informational dependencies.

• Organizations and workers can use lower information costs to reshape authority, coordination, and their positions within hierarchies.

Introduction

Last year, I came across Garicano’s idea of knowledge-based hierarchies. I had already written a post on why AI growth is not creating junior roles. At the same time, captivated by the idea that AI reduces the cost of searching for information, I developed a submission for the Google DeepMind challenge on how generative AI changes consensus formation in organizations. Since then, the technology has continued to advance, reinforcing the idea that AI reduces information search costs. However, the connection between AI and consensus-making has continued to persist in my mind. I made several attempts to launch a randomized experiment based on this idea. For now, I am describing my thinking about it.

How AI Builds Consensus

Organizations can be understood as systems designed to manage information and coordination costs. Their structures help reduce uncertainty, prevent individuals from becoming overwhelmed, and determine how information and decisions move through the organization. Historically, high coordination costs favored vertical hierarchies in which information was filtered and difficult decisions were escalated upward.

Generative artificial intelligence (GenAI), particularly when embedded in shared chat-based systems, may alter this logic. GenAI can capture knowledge from more experienced or productive workers and make it accessible to others, raising the baseline level of performance across the organization. By lowering the cost of searching for information, organizing dispersed knowledge, and reducing cognitive burdens, GenAI may also make it easier for groups to reach collective decisions or build consensus.

Organizations may therefore rely less on vertical hierarchies to manage information bottlenecks and adopt more distributed forms of coordination. This possibility is especially relevant for matrix organizations, decentralized teams, DAOs, and open-source communities, where collective decisions must often be made without strong hierarchical authority.

How AI Reduces Hierarchies 

Now, let's dig deeper to answer why the proposed mechanism may work. 

Garicano's idea of knowledge-based hierarchies helps explain why organizations divide problem-solving across different levels. No individual can efficiently acquire all the knowledge needed to solve every problem. As a result, frontline workers typically address common or routine issues, while more difficult cases are escalated to people with greater expertise.

Two costs shape the hierarchical structures: the cost of acquiring knowledge and the cost of communicating with someone who already has it. GenAI may reduce both costs at the same time. Retrieval-augmented generation systems can consolidate fragmented organizational knowledge into a shared resource, making information easier to locate (cost of acquisition) and communicate (cost of communication). Equipped with such a tool, the frontline workers reduce their dependency on the informational hierarchies. That way, GenAI may improve the speed, strategic alignment, and consensus of collective decisions. When knowledge becomes easier to acquire, informational discrepancies will be narrower, and the consensus will be more efficiently achieved in the organizational setting. As a corollary, the reduced dependency on the hierarchies to inform the organization can produce flatter organizational structures. 

The Caveat

But there is a caveat - the effect may be reversed.

When knowledge saturation and communication costs become cheaper, workers may consult experts more frequently, potentially increasing their dependence on the hierarchy even as coordination becomes more efficient. This is where the well-known Jervon's paradox comes into picture and produces produce competing prediction. 

By helping workers solve problems independently, GenAI may flatten organizational hierarchies. By making expert knowledge easier to access, however, it may also increase dependence on centralized knowledge systems and those who control them. The combined effect on organizational structure and decision quality is therefore unclear. This tension is central to understanding how organizations evolve when both knowledge acquisition and communication become more accessible.

Conclusion

GenAI, thus, has the potential to change how organizations distribute knowledge, authority, and decision-making. The direction of organizational change is, however, not predetermined.

What appears less ambiguous is that GenAI reduces the cost of finding, accessing, and applying information. This reduction creates opportunities not only for organizations but also for individual workers. Employees can use broader access to knowledge to expand their responsibilities, reduce their dependence on supervisors, and strengthen their position within organizational hierarchies. Senior specialists may similarly use AI to extend their influence across a larger number of decisions. From an entrepreneurial perspective, GenAI creates new possibilities for workers to reorganize expertise, claim new decision-making roles, and build organizational structures that were previously too costly to sustain.

The central question is therefore not simply whether AI will flatten or reinforce hierarchies. It is how organizations and workers will use lower information costs to reshape authority, coordination, and their relative positions within the organization.

References:

1. Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organization Science, 5(1), 14–37.

2. Polanyi, M. (1966). The tacit dimension. Doubleday & Company, Inc.: Garden City, New York.

Please cite this article as: 

Petryk, M. (2026, May 20). Tacit Knowledge and AI: Keeping the Human Edge in the AI Age. MariiaPetryk.com. https://www.mariiapetryk.com/blog/post-32