When employees understand how work will change through AI and that human judgment still matters, AI adoption will increase while AI anxiety will decrease.
By Stephanie Larson
“AI transformation will move at the speed of trust.”
That observation from Seramount President Subha Barry captures what many AI strategies still miss and the tension facing CHROs and senior HR leaders right now. AI is moving quickly through the enterprise, but the human strategy around it is still catching up. Without that foundation, AI becomes a change that employees are asked to absorb, not one they are equipped to shape.
When organizations undergo transformation, change must be built with people. Employees don’t need another reminder that AI is here to stay or another promise that it will catalyze productivity. They need to understand how work will change, where human judgment still matters, and how responsible use will be shaped with their participation. When AI feels pushed onto employees rather than built with them, skepticism is the result.
Many leaders may be underestimating that skepticism. According to a recent BCG survey, 76% of executives said employees feel enthusiastic and optimistic about AI adoption in their organization. And yet, when surveyed, only 31% of individual contributors expressed that same sentiment.
That perception gap matters because enterprise investment in AI continues to accelerate. Global AI spending is estimated to top $2.5 trillion this year; however, many organizations are still struggling to convert investment into measurable performance gains. According to MIT Sloan, 95% of organizations have witnessed zero measurable return on their AI investments, while only 5% of integrated pilots are extracting millions in value.
Tools can be purchased quickly. Trust cannot. Platforms can be deployed across the enterprise, but human judgment, manager readiness, role clarity, and shared understanding have to be built over time.
That distinction undergirds Seramount’s research from its report The AI Productivity Paradox. AI can collapse the cost of production, but it does not collapse the cost of judgment. Faster drafts, summaries, analyses, and recommendations still require people to verify accuracy, assess risk, contextualize decisions, and own the outcomes that follow. When organizations treat AI as a tool rollout rather than work redesign, speed can increase while trust, learning, and accountability weaken.
Many organizations are approaching AI through what Seramount calls a “speed-first” model, where output, volume, and velocity become the default measures of progress. Usage rises. Activity increases. Tools spread quickly. Judgment is expected to follow.
Human-enabled AI begins from a different relationship between people and technology. Human judgment doesn’t slow AI transformation; it’s what makes AI worth trusting. Ownership, learning, and decision quality are not signs of stalled adoption. They are the conditions that allow AI to create value.
Both approaches may use the same technologies. Only one builds the human infrastructure required for durable value.
AI is not merely a machine capability added to existing work; rather, it is unfolding inside organizations that were not designed for this speed of change or this new model of collaboration. That collaboration depends on communication, manager connection, transparent decision-making, inclusive access to learning, and clear norms for when human judgment must remain in control.
Trust is becoming the new fault line in whether organizations can make AI-enabled change legible and shared. In this new era, communication must not be defaulted to a launch email justifying the financial costs or an acceptable-use policy disseminated across Slack channels. Employees need leaders to explain the scope of what AI can accomplish and where and how it falls short. In addition, employees need a pathway for raising concerns and experimenting with the tools without being labeled resistant.
Leaders also need to be careful with language that sounds motivational in the boardroom but threatening at the team level. The familiar refrain that “AI will not replace people, but people who use AI will replace people who do not” often lands less as encouragement than as a warning: adapt quickly or become expendable. Empathetic leadership helps leaders recognize that gap between intent and impact and acknowledge that we are in a deeply transformative moment that requires vulnerability and interdependence. Unempathetic rollouts may push adoption in the short term, but ultimately, they undermine the trust and psychological safety employees need to learn, experiment, and ask questions honestly.
Trust is why managers are the hinge of AI transformation. Enterprise AI strategy becomes real in ordinary manager-employee conversations: whether a tool is safe to use, whether AI-generated work will be rewarded or questioned, whether employees can challenge an output, and whether uncertainty can be voiced without penalty.
That manager layer is already under strain. Seramount’s mental health research found that burnout is widespread among U.S. workers, with one third reporting burnout, what some leaders have identified as the consequence of being in a state of “perpetual volatility.” The strain is especially visible among managers, 80% of whom reported at least one burnout symptom.
The emotional stakes of AI adoption are becoming clearer. A 2025 Nature study found that AI adoption was associated with higher employee depression, with job insecurity and emotional exhaustion helping explain the relationship. Harvard Business Review has also warned of “AI brain fry,” the mental fatigue that can come from excessive use or oversight of AI tools, with risks including errors, decision fatigue, and intent to quit. AI adoption is not emotionally neutral; when work accelerates without clearer support, the strain lands on people first.
Managers cannot be expected to serve simultaneously as translators, troubleshooters, ethics reviewers, productivity coaches, and emotional shock absorbers without clearer support. Trust will often be determined in the manager layer, but managers need language, guardrails, escalation paths, and time to lead that work well.
The rise of AI is also making a set of humanist skills newly operational. The ability to question output, challenge assumptions, interpret ambiguity, provide context, understand human response, and connect individual concerns to structural conditions is no longer a nice-to-have. It is part of enterprise AI readiness.
That does not mean every employee needs a humanities degree. Rather, it means organizations need to build and reward the skills that make machine-human collaboration not only more trustworthy—but possible: critical interpretation, ethical reasoning, contextual judgment, perspective-taking, and communication across difference. AI can produce fluent answers. People still need to determine whether those answers are useful, fair, accurate, and appropriate for the moment.
For HR leaders, the path forward begins with trust-building and communication. That means executives must nuance both AI’s promise and its limitations. It means naming where human accountability remains nonnegotiable. It means explaining how work will change before employees are measured against new expectations. It means equipping managers to invite questions rather than merely enforce adoption. It means creating ways for employees to participate in shaping responsible use.
Speed-first AI asks people to adapt to the machine. Human-enabled AI facilitates the learning conditions that allow people and machines to change together. AI transformation will not succeed because employees are told to use new tools. It will succeed when employees understand the change, trust the people leading it, see where human judgment remains accountable, and feel invited to move with the organization into what comes next. 
Stephanie Larson is principal of strategic research for Seramount. 
Yes! I would like to receive new content and updates.
Leave a Reply