AI-driven physiotherapy risks hollowing out routine clinical roles and professional authority while shifting value toward tech platforms; preserving patient outcomes will require revaluing therapists' relational and complex‑case skills and redesigning roles rather than protectionism.

Main Finding

The paper argues that physiotherapy — and orthodox health professions more broadly — are entering a post-professional era driven by commodification, political delegitimation of professional monopoly, and digitally mediated automation. Extending Deleuze’s “societies of control,” the author proposes a new phase: a “society of indifference,” where predictive algorithms replace individual choice, offering personalized care without requiring active patient decision‑making. Using the UK AI physiotherapy clinic Flok Health as a case study, the paper shows that end‑to‑end AI care threatens not only routine tasks but practitioners’ agency and the institutional structures that sustained professionalized rehabilitation. The author concludes that preserving outcomes will require reconfiguring what is shared or preserved of physiotherapy’s core “intensities” rather than defending traditional professional turf.

Key Points

• Three concurrent forces erode professional physiotherapy:
  • Atomization/commodification: bodies and interventions are unbundled into discrete, marketable tasks under late capitalism.
  • Political delegitimation: professional claims to moral authority and unique expertise are undermined across ideological lines.
  • Technological disruption: digital platforms and AI automate assessment, diagnosis, treatment plans, and monitoring, hollowing jurisdictional claims.
• Deleuze’s “societies of control” is a productive theoretical lens but must be updated for generative AI and pervasive predictive systems.
• “Society of indifference”: a proposed post‑neoliberal regime in which algorithms make (and normalize) choices for users, delivering personalized services while diminishing individual deliberation and accountability.
• Flok Health exemplifies the era: AI-driven triage, assessment, exercise prescription, and discharge for musculoskeletal conditions reduce demand for human-led, iterative, judgment‑based physiotherapy.
• Risks identified:
  • Loss of practitioner agency and professional identity.
  • Potential gaps in care for complex or psychosocially mediated rehabilitation needs that AI or consumer choice alone may not meet.
  • Institutional erosion: training pathways, norms of professional accountability, and organizational niches may shrink or transform.
• Recommended response: instead of protectionism, physiotherapy should identify, articulate, and share its core “intensities” (skills, relational capacities, embodied knowledge) and adapt roles where human‑AI complementarities are strongest.

Data & Methods

• The paper is primarily conceptual and theoretical, synthesizing:
  • Deleuzian social theory on societies of control.
  • Political economy analyses of professionalization and neoliberal restructuring.
  • A qualitative case study of Flok Health (UK’s first AI physiotherapy clinic) used illustratively to show how end‑to‑end AI threatens clinical jurisdiction.
• Methods likely include:
  • Discursive and institutional analysis of professional claims and market trends.
  • Case-based description of technological capabilities and clinical workflows at Flok Health.
  • Normative reflection on sociotechnical futures and professional practice.
• Limitations:
  • Little or no large‑scale quantitative evidence on clinical outcomes, labor market impacts, or patient preferences is presented.
  • Generalizability beyond musculoskeletal physiotherapy and the UK context is suggested rather than empirically established.
  • Empirical claims about future trajectories rest on theoretical projection and a single illustrative case.

Implications for AI Economics

• Labor demand and task reallocation
  • Short‑run: routine, codifiable physiotherapy tasks face high automation susceptibility; demand for such tasks (and associated junior roles) may fall.
  • Long‑run: potential for job redesign — fewer full‑service clinicians, more hybrid roles (AI supervisors, complex-case specialists, rehabilitation coordinators).
  • Wage effects depend on complementarities: high-skilled relational tasks may see wage premia, while commodified tasks compress wages.
• Professional rents and market structure
  • Professional monopoly rents are eroded as AI lowers entry costs and unbundles services; market concentration may shift toward tech/platform firms with scale economies rather than clinician collectives.
  • Platformization can create winner‑take‑most dynamics; AI providers could capture scale-based rents (data, models, distribution).
• Consumer choice, agency, and demand elasticity
  • “Society of indifference” implies reduced active choice and potentially lower consumer search costs, changing price sensitivity and perceived value.
  • Algorithmic paternalism could increase uptake of recommended protocols (raising short‑term adherence) but may reduce consumer learning and long‑term engagement.
• Welfare tradeoffs
  • Efficiency gains (lower unit costs, broader access) vs. losses in non‑technical care dimensions (empathy, individualized judgment, complex reasoning).
  • Potential coverage gains for low‑complexity conditions; risks of underservice for complex or socially embedded rehabilitation needs.
  • Distributional concerns: differential access to high‑quality human care may exacerbate inequalities if AI becomes the default low‑cost option.
• Incentives, regulation, and accountability
  • Existing professional regulation is ill‑fitted for autonomous AI: need new frameworks for accountability, certification, and safety of algorithmic treatment.
  • Payment systems (fee‑for‑service vs. bundled payments) will shape incentives for AI substitution versus human complementarity.
  • Data governance and model transparency will become central economic policy levers.
• Research and measurement priorities
  • Empirical evaluation of clinical outcomes, cost‑effectiveness, and distributional impacts of AI physiotherapy vs. human care.
  • Task‑level mapping to identify which physiotherapy components are automatable and which are complementary.
  • Labor market studies tracking wage dynamics, occupational transitions, and training pathways for allied health professions.
  • Market structure analysis of platform providers, data network effects, and barriers to entry for clinician‑led alternatives.
• Policy and strategic responses for stakeholders
  • Regulators: develop certification, continuous auditing, and liability regimes tailored to AI clinical agents; ensure pathways for escalation to human clinicians.
  • Payers: design reimbursement that values humanly delivered intensities (complex reasoning, relational continuity) and avoids perverse incentives to offload complex cases to under‑resourced settings.
  • Professional bodies and training institutions: pivot toward teaching hybrid skill sets (AI oversight, triage of complexity, interpersonal rehabilitation skills) and formalize new roles.
  • Firms: explore human‑AI complementarities (e.g., AI handling routine monitoring, clinicians focusing on complex, relational, and supervisory tasks).
• Broader economic outlook
  • The case illustrates a wider pattern: professional services across sectors face replacement of codifiable tasks by AI, with implications for labor market composition, welfare, and institutional design.
  • Policy choices will determine whether AI-driven efficiency yields broadly shared gains or concentrates benefits while hollowing out capacities needed for complex human needs.

Suggested next empirical steps for economists: randomized trials comparing AI-only, human-only, and hybrid care for outcomes and costs; matched employer–employee studies tracking occupational transitions; market analyses of platform emergence and data‑driven concentration in health services.