The Efficiency Trap: Cognitive Load, Skill Atrophy, and Critical Thinking

Session Overview

AI has become a genuine thought partner in how work gets done. It helps people move faster, produce cleaner outputs, and tackle tasks that once required far more effort. But there is a tension that most organizations are not naming clearly enough: efficiency and learning are not always compatible. Learning is, by nature, messy. Mastery takes time. And when AI removes the friction before people even notice it was there, we may be quietly eroding the very skills and judgment we are counting on.

At Transform 2026, I joined moderator Naomi Titleman (Co-Founder, future foHRward), Shweta Vohra (Chief People Officer, GitHub/Microsoft), and Greg Karanastasis (SVP & Head of Talent & Development, CVS Health) for a 30-minute panel on this exact tension: the efficiency trap, cognitive load, skill atrophy, and what it takes to protect critical thinking in an AI-enabled workplace.

Psychological Readiness as an Upstream Strategy

Most organizations are approaching AI like a technical rollout. At Intermountain Health, we have taken a different approach. We treat it as a human readiness challenge.

In healthcare, we do not wait for disease to show up before we intervene. We focus upstream. We have applied that same thinking to AI adoption. Instead of waiting until people feel behind or overwhelmed, we build psychological readiness before the moment of need.

That means we are not pulling people into week-long trainings or expecting a single AI literacy event to stick. We are embedding learning into the rhythm of their day. Small, consistent exposure — micro-learnings, tip-of-the-day delivery, practical use cases tied to real workflows — builds confidence over time. And when confidence goes up, adoption follows. When confidence stays low, even the best tools go unused, or worse, get used poorly.

The goal is not AI literacy on a certificate. It is practiced confidence — the kind that shows up when the stakes are real.

Human-at-the-Helm: Protecting Productive Struggle

There is a real paradox at the center of this conversation. AI removes friction, but friction is how we learn. The pressure to be efficient is real and legitimate — but if every difficult moment gets handed off to a tool, we never build the muscle.

We talk a lot in the industry about keeping humans in the loop. At Intermountain, we think about it differently: human-at-the-helm. Use the governing frontal lobe we have all been given. Make sure human judgment remains central — not as a checkpoint at the end, but as the guide throughout.

There are domains where we actively want productive struggle to continue. First-draft thinking. Clinical reasoning. Strategic problem framing. Performance conversations. These are the places where the process of working through something is itself the point. If AI writes your first draft every time, you never develop the craft. We want AI to accelerate clarity — not replace cognition.

At scale, this becomes less about rules and more about norms. We coach leaders to ask a simple question: Did AI help your thinking, or did it do your thinking? That distinction, held consistently, is how culture changes.

Passive vs. Intentional: The Accountability Gap

There is a lot of energy in the HR field around resistance to AI — people who are hesitant, skeptical, or slow to adopt. But there is an equally important problem at the other end of the spectrum: people who are too comfortable. They take the AI output, hit send, and move on. No triangulation. No judgment. No real engagement with the work.

Passive AI use feels productive. The output looks polished. But nothing was learned, and no one is accountable for the thinking behind it. Intentional AI use is different. It means layering your expertise on top of the output, questioning what came back, and making the final call yours.

Leaders have to model that difference visibly. At our People Team offsite, leaders and their teams engaged directly with AI through a structured deep dive using real People Team use cases. One function extended the work by collaboratively building an AI agent, with leaders learning alongside their teams — not delegating the experimentation to someone else. That is what intentional adoption looks like in practice.

Building a culture of accountability around AI is not primarily a technology problem. It is a leadership problem. Who is setting the expectation that AI is a starting point, not a shortcut? Who is modeling critical review of AI outputs — out loud, where teams can see it? That is the work.

Perspectives Worth Carrying from the Panel

Greg Karanastasis described CVS Health's response to skill atrophy as a reskilling strategy built around the shift from writing to editing. Their CVS AI Academy delivers executive-level three-hour deep dives alongside AI Week experiences for all employees. Notably, "Champions AI" has become a formal leadership competency — a signal that AI fluency is now a built expectation of leaders, not an optional interest.

Shweta Vohra offered a reframe that I think cuts right to the heart of the issue: the concern is not skill atrophy — it is skill activation. The question is not what skills are disappearing, but which new or underleveraged human skills AI should be freeing up capacity to develop. Judgment, synthesis, empathy, complex communication — these are not diminished by AI. They should be more central than ever. The organizations that get this right will give people permission and guardrails to experiment, then share what works and what does not. The cost of inaction is higher than the cost of imperfection.

Practical Takeaways

1. Start with one habit, not AI training: Pick one moment in your day and ask — am I using AI to think better, or to think less? Then talk about it with your team. Culture changes through conversations, not through tools.
2. Build psychological readiness upstream: Do not wait for resistance or overwhelm to arrive. Embed small, consistent learning touchpoints into the flow of work before people feel behind.
3. Protect productive struggle intentionally: Identify the domains in your organization where the process of working through something is itself the development. Be explicit about where AI should augment thinking and where it should not replace it.
4. Coach the passive-to-intentional shift: Train leaders to model critical review of AI outputs out loud. Give teams a common language for the difference between using AI to accelerate thinking and outsourcing it entirely.
5. Treat adoption as ongoing experimentation: The organizations that lead in this era are not the ones who got the playbook right first. They are the ones who started, learned fast, and kept going — with clear outcomes, risk boundaries, and role clarity.

Related Insights

• Future-Ready HR - Driving Workforce Transformation Through Technology and Inclusion Shared focus: moving from experimentation to repeatable execution.
• The Agent-Powered Enterprise Shared focus: governance, accountability, and human oversight in AI-enabled work.
• Building AI-Ready Workforces in Healthcare and Pharma Shared focus: workforce readiness as the multiplier for AI value.

Related Event

Visit the official event page for agenda and speaker details.

Return to Appearances on lancebradshaw.com to view the full speaking portfolio timeline.