Divya van Mahajan

The New Jobs AI Will Create

ai-daily-brief-podcast

Overview

This episode of the AI Daily Brief (hosted by an unnamed presenter) argues that the dominant narrative around AI and jobs — that AI will destroy white-collar work — relies on a flawed hidden assumption. The presenter contends that the field has spent enormous energy cataloguing jobs at risk while doing almost no work to identify the new jobs AI will create. The talk attempts to correct that imbalance by building a first-principles framework for where new demand and new roles will emerge in an AI-enabled economy, using healthcare as a detailed case study.

Source video: No URL was provided for this episode.

Prerequisites

  • Basic understanding of labour economics (supply and demand, price elasticity)
  • Familiarity with the current public debate around AI and employment
  • General awareness of how AI agents and large language models are being deployed in professional services
  • No domain expertise in healthcare is required (the presenter explicitly disclaims it)

Main Points

1. The Hidden Assumption in AI Job Apocalypse Narratives

  • Most AI-and-jobs analysis treats AI purely as a labour supply story: AI increases labour supply → labour gets cheaper → workers are displaced.
  • This analysis implicitly holds demand constant — a version of the “lump of labour fallacy.”
  • Historically, the assumption that demand stays constant as labour supply increases has never held true.
  • The more productive question is not which jobs disappear but where will demand expand.

2. Six Categories of Demand Elasticity

The presenter identifies six ways demand can expand as AI lowers the cost, complexity, and friction of producing goods and services:

  • Price elasticity — “I wanted it but it cost too much.” AI lowers the cost floor and activates new buyers.
  • Access elasticity — “I wanted it but couldn’t get it.” AI reduces provider scarcity, wait times, geographic barriers, and institutional bottlenecks.
  • Complexity elasticity — “I needed it but the system was too confusing.” AI makes opaque systems (tax, insurance, immigration, medical recommendations) navigable.
  • Continuity elasticity — “I get help occasionally but would benefit from help all the time.” AI makes always-on monitoring and support operationally viable at scale.
  • Personalization elasticity — “I get the generic version but would value something made for me.” AI lowers the cost of customisation.
  • Relational/value elasticity — “I want this service to be more human, meaningful, or trusted.” Drawing on economist Alex Emas’s concept of the relational sector, where the human provenance of a service is integral to its value.

3. Two Types of Demand Unlock

  • The Affordability Unlock — The same menu at a lower price. Existing services reach entirely new buyers who were previously priced out. Example: a small business that could never afford a $5,000 design project becomes a buyer when AI brings that cost to $500, activating a vast long-tail market of first-time agency clients.
  • The Possibility Unlock — A new menu entirely. AI makes service models operationally viable that could not previously exist at scale, creating net-new demand for things that no one was demanding because they did not yet exist. Example: continuous preventative healthcare — no one currently demands it as a category because it has never been viable.

4. Addressing the AGI Objection

  • The most common counter-argument: “Won’t AGI just eat those new jobs too?”
  • The presenter argues this question is framed incorrectly. It asks only can AI perform the task? (a capability question) rather than does AI-only delivery satisfy the demand? (a service design question).
  • Many roles exist not because of capability gaps but because of how a service must be delivered — trust, accountability, presence, and relationships are part of the value, not incidental to it.

5. The Seven Categories of Human Premium

The human premium is defined as the portion of economic value that remains attached to human involvement even when AI can perform the underlying task:

  • Relational — “I want this delivered by someone who knows me.” Continuity, memory, and accumulated trust are integral to the experience.
  • Embodied presence — “I want someone there with me.” Physical presence is part of the value (e.g., a nurse in the room, a trainer correcting form).
  • Trust — “I want to talk to a person before I act.” Social proof and human validation make AI-generated recommendations emotionally acceptable and actionable.
  • Accountability — “Someone has to own this.” People want a human who signs off, escalates, explains, and bears responsibility when things go wrong.
  • Translation — “I don’t know how to ask for what I need.” Even with AGI, there is economic value in humans who can turn messy intent into usable AI-mediated work. The margin on top of the underlying AI cost-of-goods-sold does not disappear.
  • Behaviour change — “I know what to do but need help doing it.” For certain domains (fitness, health habits), humans are more effective accountability partners than AI.
  • Provenance and status — “A human made it is part of why it matters.” Arts, crafts, live performance, and bespoke services embed the human signature as a product feature.

6. Healthcare Case Study — Three Net-New Roles

The presenter uses continuous preventative healthcare as a concrete example of the possibility unlock, proposing three illustrative roles:

Role 1: Continuous Care Navigator

  • Human layer between patient and AI monitoring system.
  • AI handles: data ingestion from devices/labs/pharmacies, baseline establishment, deviation detection, urgency ranking, documentation, escalation routing.
  • Human handles: reviewing flagged cases, calling patients, noticing fear/shame/avoidance, coordinating clinician escalation, closing the loop.
  • Human premiums: trust, accountability, translation, behaviour change, relationship.
  • Estimated scale: 276,000–1,200,000 jobs (conservative: 40M enrolled patients at 150:1 ratio; aggressive: 120M enrolled patients at 100:1 ratio).

Role 2: Care Plan Outcome Specialist

  • Owns the gap between medical advice and real-world execution.
  • AI handles: tracking care plan milestones, medication adherence, labs, screenings, integration into continuous care record.
  • Human handles: discussing why the plan is not working, solving practical barriers (cost, transport, fear, family dynamics).
  • Scale: potentially hundreds of thousands of net-new positions.

Role 3: Health Data Operations Specialist

  • Owns reliability, integration, governance, and clinical usability of the data layer.
  • AI handles: pulling and normalising data, flagging anomalies, generating pipeline diagnostics and audit logs.
  • Human handles: permissions, consent, audit governance, resolving device/integration failures, translating between clinical and IT requirements, institutional accountability.
  • Scale: one specialist per 1,000–2,000 patients; estimated 20,000–40,000 net-new roles.

7. Broader Sector Patterns and Six Families of New Roles

The presenter argues the same logic applies across sectors — small business professional services, education, mental health, personal finance, elder/child care — and distils the new job landscape into six broad role families:

  • Navigators — Help people enter and move through systems too complex to face alone.
  • Continuous support workers — Provide ongoing human support around AI-monitored systems.
  • AI-augmented service operators — Use AI to deliver cheaper professional services to previously unreachable market segments.
  • Data and operations specialists — Make AI-enabled service models reliable within real institutional systems.
  • QA, safety, and compliance roles — Ensure AI-mediated services are safe, auditable, legal, and fair.
  • Escalation specialists — Handle the hardest cases that AI routes upward.

Key Concepts

  • Lump of labour fallacy — The erroneous belief that there is a fixed amount of work to be done, so technology reducing the labour required for some tasks necessarily leaves workers unemployed.
  • Demand elasticity — The degree to which demand for a good or service expands in response to changes in price, access, complexity, continuity, personalisation, or relational value.
  • Affordability unlock — AI-driven cost reduction that brings existing services within reach of buyers who were previously priced out, activating long-tail markets.
  • Possibility unlock — AI making entirely new service models operationally viable for the first time, creating net-new categories of demand.
  • Human premium — The portion of economic value that remains tied to human involvement even when AI can perform the underlying task; the presenter’s answer to the AGI objection.
  • Relational sector — Term from economist Alex Emas describing services in which the human provenance and mode of delivery are intrinsic components of the product’s value.
  • Continuous preventative care — A hypothesised AI-enabled healthcare model that is always-on, data-driven, personalised, and proactive, as opposed to the current episodic reactive paradigm.
  • Continuous care navigator — A proposed new role: the human coordination and escalation layer sitting atop an AI patient-monitoring system.
  • Care plan outcome specialist — A proposed new role: the human responsible for bridging the gap between medical recommendations and patient execution.
  • Health data operations specialist — A proposed new role: the human accountable for the governance, reliability, and clinical usability of the data infrastructure underpinning AI-enabled care.

Summary

The presenter’s central argument is that the AI-and-jobs debate is systematically incomplete because it analyses AI only as a labour supply shock while ignoring demand. Demand is not fixed: it expands through at least six distinct elasticities — price, access, complexity, continuity, personalisation, and relational value — and AI’s ability to lower costs and remove friction will trigger both an affordability unlock (existing services reaching new buyers) and a possibility unlock (entirely new service models becoming viable). The objection that AGI will simply consume any new jobs created is answered by the concept of the human premium — seven categories of value (relationship, embodied presence, trust, accountability, translation, behaviour change, and provenance) that do not automatically transfer to AI delivery even when AI can perform the underlying task. Grounding the argument in healthcare, the presenter sketches three plausible net-new roles — continuous care navigator, care plan outcome specialist, and health data operations specialist — and estimates they could represent hundreds of thousands to over a million positions in the United States alone. Generalising across sectors, the talk concludes that AI will generate six broad families of new roles — navigators, continuous support workers, AI-augmented service operators, data and ops specialists, QA/safety/compliance roles, and escalation specialists — and that while short-term disruption is real and should not be minimised, the longer-term trajectory is one of significantly expanded human work in an economy whose demand frontier AI has enlarged.