Leadership Trust and AI: Why the Human Layer Decides Adoption

The most common hurdle for executives is not a technology hurdle. In a recent appearance on Signal Room, Chris mentioned studies that reported less than a quarter of employees trust their CEO. The guest that episode, Diane Weaver, the co-founder of an organization developing human-centric AI for the workplace, and a self-described systems thinker, did not consider this a peripheral concern. For her, this defines the precondition that dictates whether any AI project is successfully implemented.

Weaver's main point reframes the debate that is most commonly misplaced. Most of the discussions around trust and AI concern the level of trust in the AI model. For example, does the AI model produce reliable outputs? Does the AI model hallucinate? In fact, she almost immediately set aside the question of trust in AI models. She stated that the key barrier to the success of any AI initiative, and the relevant trust in the organization, is trust of the leaders of the organization who are implementing the AI systems and the employees of the organization. This is the layer that AI has the potential to strengthen or undermine.

The Key Trust Issue Is Trust Among Humans, Not Trust in Technology

Weaver argued that this issue of trust resides more in the organization than where the technology resides. Trust in the technology will be achieved once the AI model is validated, its constraints are documented, and safeguards are established to contain the technology. Weaver's main point is that the organization will have the best AI technology at its disposal, but the AI initiative will fail if the employees who are required to work with the technology do not trust the leadership of the organization.

She relied on one definition to explain how trust is built. Trust develops because people do what they say. People decide whether to trust a new statement by a leader based on evidence of the leader's past. A definition of trust is simple, and somewhat boring, but saying a leadership team follows up on an announcement that an organization is about to transform its AI captures the essence of the promise. The promise is kept only by action. Trust in the promise depends on the leadership team keeping promises in the past.

Why AI Lands on Top of an Existing Trust Deficit

Weaver discusses a phenomenon that many are familiar with as a technology requirement progresses through an organization. AI is no exception: adoption starts at the leadership level and then comes down the line, and the gap between that leadership level and the frontline reality of getting the work done is substantial. Most people on the front line are completely exhausted by ever-rising productivity demands. To reimagine the work, how each functions in the context of new technology, will further strain the already failing leadership floors. The AI initiative arrives on a relationship deficit. The tech is the visible object of trust that is not there.

The intersection between leadership and trust diverges from the influence of an organization's culture and AI readiness. It is not simply whether an organization's culture is curious or resistant. It is whether the relationship between those who set mandates and those who execute them is robust enough to endure further change.

The Pitfall of Leadership

Weaver points out the conundrum of a leadership position in a more straightforward manner. Once a person is appointed to a leadership position, they realize how restricted their options are. The constraints of the position are invisible to those who are outside the position. Because there is no precedent of changes in the position, it makes it difficult to communicate to the organization something new or different. Suggesting change without precedent causes uncertainty rather than confidence.

The phenomenon of leadership is the articulation of things that cannot actually be accomplished. Each failure to accomplish a commitment is a withdrawal from an already low trust account. The effect of this is particularly damaging on an AI program because AI initiatives are inherently protracted, uncertain, and iterative. A leader who commits to a quicker than expected AI program, fewer personnel, or smoother transitions, erodes trust they will need when the program is inevitably stalled.

The more difficult stance is to describe the limitations of what you are prepared to offer. It is a form of follow-through to describe the limits of what you can offer, because it describes the limits of what you may commit to.

Disengagement Is Not a Frontline Problem

It would be ideal for workplace disengagement to be a problem solely for those furthest away from the centers of power. Weaver, who appeared on the show, said that about two-thirds of the global workforce is reported to be disengaged, and she was cautious in her explanation of what it means. The most telling thing for Weaver, and in her view, the most troubling part of the report, is that disengagement is found at all levels of the workplace, from the most junior employees to the most senior leaders.

Her theory is that disengagement is a function of the misalignment of expectations. She describes the disengagement of employees in several dimensions. There is the gap between what a person actually wants to be doing and what the person spends the majority of their time doing. There is also a gap between what a worker deems to be a valuable contribution and the goals employees are expected to meet, and the targets and valuable contribution of the job are at cross-purposes. And there is the most insidious gap of all, which is a worker achieving the goals the business has set for them while those goals do not meaningfully contribute to the business in any way.

Before using AI to increase productivity, leaders should consider this diagnosis. Misalignment of tasks can be a source of disengagement, and automating those tasks does not solve this problem. Instead, it can intensify disengagement and drive employees to work even more misaligned.

The Steward Mindset Keeps Humans at the Helm

Weaver recommends a shift in position. It is important for each individual to evolve and be (or at least act) as a steward of technology, rather than a user. Technology users are given tools and taught how to operate them. In contrast, a steward of technology understands the tool's purpose, identifies its shortcomings, and evaluates its use. This is the distinction that keeps the human at the helm rather than along for the ride.

She said it is important to understand why this is happening now. According to her, a regular language model is designed to maintain engagement by constantly asking deeper and deeper follow-up questions and extending the conversation. The technology that we employ alters who we are; if we spend considerable time with a system designed to optimize engagement, we begin to speak and behave in alignment with that system. The steward's defense is the set of durable human skills she kept naming — questioning, critical thinking, empathy, reading a room, and the judgment to think through the impact of an action. These are the skills that leaders should develop in their workforce, rather than relying on technology.

How Hutchins Approaches Leadership Trust and AI

We meet this condition constantly: a technically sound AI initiative that stalls because the human layer underneath it was never addressed. Our work treats trust and capability as part of the deployment, not a soft afterthought to it — building the AI literacy that lets teams act as stewards rather than passive users, and tying any AI rollout to the responsible governance that gives leaders something concrete to stand behind when they ask people to trust the change. The aim is to close the distance between what leadership commits to and what the front line experiences, because that distance is where adoption quietly dies. These themes run throughout The Signal Room podcast, where guests working at the intersection of AI and the workforce describe what it takes to keep humans at the center of the change.

Authoritative sources

• WHO — Ethics & governance of AI for health
• NIST AI Risk Management Framework
• Agency for Healthcare Research and Quality
• HIMSS
• U.S. Department of Health & Human Services — HIPAA