The dominant conversation about AI and work asks the wrong question. Rather than “How many jobs will AI eliminate?” we need to ask: “What conditions allow human beings to move toward health, coherence, and flourishing in AI-augmented environments?”
Working Paper · April 2026
21-minute read · Request a free copy
Something important is happening in American workplaces, and most organizations are not equipped to see it clearly. Artificial intelligence has moved from specialized research domain to core infrastructure of work — authoring reports, diagnosing patients, writing code, screening candidates — in fewer than three years. The pace of this transformation outstrips any comparable technological shift in modern industrial history.
Our response to it has been almost entirely framed in pathogenic terms: job loss projections, displacement anxiety, surveillance capitalism, algorithmic management. These are legitimate concerns. But they illuminate the disease without illuminating the conditions for health. They identify the threat without mapping the path toward resilience.
“The question is not whether AI will change work — it already has. The question is whether that change moves workers toward coherence or toward chaos.”
— Institute for Salutogenesis, Working Paper 2026
The Institute for Salutogenesis proposes a different starting orientation — one rooted in Aaron Antonovsky’s foundational insight that health is not the absence of stressors, but the presence of resources that enable navigation of a stressor-saturated world. Our new working paper applies this lens systematically to the AI transition.
THE EVIDENCE DEMANDS ATTENTION
In March 2026, the BCG Henderson Institute published what may be the most consequential occupational health study of the decade. Surveying 1,488 full-time U.S. workers, researchers identified a new clinical phenomenon they named “AI brain fry” — cognitive overload produced not by the use of AI, but by the sustained mental effort of monitoring, evaluating, correcting, and managing AI-generated outputs at a pace and complexity that exceeds the brain’s processing capacity.
14%. of AI-using workers experience brain fry
(26% in marketing roles)
39%. higher intent to quit among affected workers — a direct talent retention crisis
40%. of AI efficiency gains lost to correcting, rewriting, and fact-checking outputs
Productivity, the study found, peaks at three simultaneous AI tools. Adding a fourth causes measurable decline. The researchers also found that workers who used AI to replace routine, repetitive tasks reported less burnout — not more. The technology itself is not the variable. How organizations deploy it is.
Alongside brain fry, a deeper and more insidious risk has emerged: deskilling. Research published in AI & Society documents what one team calls “capacity-hostile environments” — settings where AI mediation systematically impedes human capacity cultivation. In healthcare, AI dependence has been shown to diminish diagnostic reasoning, reduce tacit clinical knowledge, and weaken ethical sensitivity. In knowledge professions broadly, AI dependence erodes activity awareness and competence maintenance. Workers may not notice the erosion until the capacity is needed and absent — a catastrophic combination in high-stakes fields.
THE SALUTOGENIC REFRAMING
Aaron Antonovsky’s Sense of Coherence (SOC) identifies three dimensions of the orientation that enables human beings to navigate a stressor-saturated world toward health:
THE SENSE OF COHERENCE — APPLIED TO AI-AUGMENTED WORK
Comprehensibility
Workers understand what AI systems do, why they make the decisions they make, and how their role fits the new workflow
Manageability
Workers have adequate training, tool limits, time, and authority to handle AI-augmented tasks without being overwhelmed
Meaningfulness
Workers experience their AI-augmented role as genuinely valuable — their uniquely human contributions are real and recognized
Each of these dimensions is under acute threat in poorly designed AI implementations. Comprehensibility is disrupted by opaque, rapidly-changing AI tools. Manageability is violated by the cognitive overload of brain fry. Meaningfulness is undermined when professional identity is organized around tasks AI now performs faster and at scale.
But here is the central thesis of our working paper: the same technology producing pathogenic effects in some organizational contexts has the potential to function as a Generalized Resistance Resource in others — a tool that strengthens, rather than erodes, workers’ capacity to navigate complexity and sustain coherent engagement with their work. The difference is design, not destiny.
“A salutogenic organization is not one that eliminates the challenges of the AI transition. It is one that builds — in every worker, every team, and every system — the coherence required to navigate those challenges toward health.”
— Institute for Salutogenesis, Working Paper 2026
WHAT THE WORKING PAPER CONTAINS
Our full working paper — available for free download below — develops this framework across eight substantive sections and a practical organizational assessment tool:
The psychology of work in pre-AI baseline conditions: Self-Determination Theory, the Job Demands-Resources model, and the role of craft identity in occupational well-being
The architecture of AI augmentation — from generative to agentic AI — and what that transition means for cognitive load and worker agency
A detailed analysis of AI brain fry, deskilling, and occupational identity erosion, grounded in the latest empirical evidence
Salutogenic reframing: how to assess whether a given AI implementation strengthens or undermines workers’ Sense of Coherence
Job-category analysis across four representative sectors: healthcare professionals, knowledge workers, customer-facing and service workers, and creative professionals — with specific AI impact patterns and salutogenic strategies for each
Strategic recommendations at individual, organizational, and systemic levels
The Salutogenic AI Workplace Assessment (SAWA): a 21-item diagnostic instrument organizations can use to evaluate their AI implementation against all three SOC dimensions
WHO THIS IS FOR
This paper is written for organizational leaders who sense that something important is being missed in their AI implementation conversations; for human resources and occupational health professionals looking for a coherent framework that goes beyond stress management and productivity metrics; for clinicians, educators, and knowledge workers trying to make sense of what is happening to their professional identities; and for policymakers and researchers working on the human dimensions of the AI transition.
It is grounded in empirical research. It is oriented toward practical action. And it is written from a conviction that how we navigate this transition will determine not just organizational performance, but the quality of human working life for a generation.
A NOTE ON OUR APPROACH
The Institute for Salutogenesis works from Antonovsky’s foundational principle that health promotion requires more than the elimination of pathogenic conditions — it requires the deliberate construction of health-generating ones. In the context of AI and work, this means we are not primarily interested in cataloguing harms. We are interested in identifying the organizational conditions, individual capacities, and design principles that enable workers to experience their AI-augmented work as comprehensible, manageable, and meaningful. That is the salutogenic ambition. We believe it is achievable. We believe the stakes of failing to pursue it are far higher than conventional workforce analysis has yet recognized.
