Human Premium Assessment
Score seven irreplaceable human value categories to determine which roles survive AGI
The Human Premium Assessment separates the capability question—can AI do this task?—from the service design question—does AI-only delivery actually satisfy the demand? It identifies seven categories of value that do not automatically transfer when a human is removed: Relationship, Embodied Presence, Trust, Accountability, Translation, Behavior Change, and Provenance. By scoring each premium for a given role or service, users determine which roles are structurally durable versus merely transitional. Roles protected by at least two strong premiums survive compounding AGI capability; roles protected only by current capability gaps are ultimately at risk. The framework gives practitioners a principled, specific answer to the question: won't AGI just eat those new jobs too?
- AI can eat tasks but does not automatically eat demand for trust, accountability, or human relationship
- The relevant question is service design, not just capability: does AI-only delivery satisfy what the market actually wants?
- Human premium categories are independent; a role may carry one strong premium or several
- Premiums grounded in legal liability, physical necessity, or deeply embedded market expectations are structurally durable
- Premiums based only on current unfamiliarity with AI are transitional and will erode
- Roles designed explicitly around durable human premium categories create competitive positioning that AGI capability gains cannot easily undercut
- Inventory all tasks AI can theoretically perform in the roleList every task the role currently performs and mark which ones AI can do today or plausibly within a five-year horizon. Be ruthless and assume AI capability will continue compounding. This step forces an honest reckoning with which parts of the role are genuinely at risk before any protective rationalization sets in.Pro tipUse current AI tools to actually attempt each task. Tasks you assumed were too nuanced often turn out to be AI-doable; tasks you assumed were simple sometimes reveal surprising human dependencies.WarningDo not list only the obvious operational tasks. Include soft activities like explaining recommendations, managing client emotion, and handling exceptions—these are where premiums most often hide.
- Map each task to the seven Human Premium categoriesFor each task, ask whether removing the human from that specific activity would decrease demand for the overall service. If yes, identify which premium applies: Relationship (continuity and accumulated trust); Embodied Presence (physical being there); Trust (social proof and human validation); Accountability (legal or moral ownership); Translation (turning messy desire into usable AI input); Behavior Change (human accountability for personal change); Provenance (human-made as integral to value).Pro tipA single task can carry multiple premiums simultaneously. A nurse's bedside visit carries both Embodied Presence and Trust, making that role component doubly durable.
- Score each premium on structural durabilityRate each identified premium from 1 (based on current habit or unfamiliarity with AI, likely to erode within years) to 5 (structural: rooted in legal liability, physical necessity, or deeply embedded market expectation). Scores of 4–5 indicate premiums unlikely to erode even as AI capability compounds significantly.Pro tipAccountability premiums backed by legal structures such as licensure, malpractice, and fiduciary duty tend to score 5 because weakening them requires regulatory action, not just improving AI.WarningBe honest about premiums that feel strong today but are really transitional. 'People prefer humans because they're not used to AI yet' is a score of 1 or 2, not 4 or 5.
- Identify the role's durable core from its highest-scoring premiumsSelect the two to three tasks with the highest-scoring premiums—these form the durable core of the role. Everything else is either automatable now or will become so. The durable core is what the role should be designed and hired around going forward, not the full current job description.Pro tipIf the role has fewer than two tasks with premiums scoring 4 or above, the role as currently designed may be transitional; consider whether it can be redesigned around higher-premium activities or consolidated with another role.WarningAvoid protecting the entire current job description by scoring every task highly. The goal is honest identification of what is genuinely durable, not rationalization of the status quo.
- Redesign the role around its durable premiumsRewrite the role's responsibilities so high-premium tasks are the primary focus and AI handles low-premium tasks. This typically means the human concentrates on judgment, relationship management, escalation, and accountability while AI handles data processing, summarization, scheduling, and routine communication.Pro tipThe redesigned role often handles a larger caseload than before because AI absorbs volume work. Build this increased throughput into capacity planning and compensation models.WarningDo not eliminate AI-assisted tasks from the job description entirely. Humans still need to supervise, audit, and take accountability for AI outputs as part of their Accountability premium.
- Position the service by naming the human premium explicitly to buyersCommunicate to buyers precisely which human premium they are purchasing. Customers who understand they are paying for Accountability, Relationship, or Translation are far more willing to pay a premium and far less likely to defect to AI-only options once they understand what they would lose.Pro tipName the premium concretely in marketing: 'a licensed professional who is legally accountable for your plan and knows your full history' is more compelling and defensible than generic 'personalized service'.
As AI handles data ingestion, baseline establishment, anomaly detection, and documentation in healthcare, a Continuous Care Navigator role emerges as the human layer above that system. Its durable premiums are Trust (patients want a person to validate AI flags), Accountability (someone owns the escalation decision), Behavior Change (humans respond to a person calling them differently than to an automated alert), and Relationship (continuity of care matters emotionally and clinically). All four premiums score 4 or 5 structurally.
A freelancer providing design, legal review, or marketing to small businesses carries a Translation premium scoring 5: the small business owner lacks time, vocabulary, and context to prompt AI tools well enough to get usable, workflow-integrated outputs. Even if AGI compresses the base production cost to near zero, economic margin remains for a human who converts messy business needs into AI-mediated deliverables and integrates them into the owner's actual operations.
An AI app can generate optimal training programs, track progress, and adjust loads—surpassing average trainer programming knowledge. Yet the Behavior Change premium scores 5 for large market segments: people follow a human trainer's instruction and abandon AI apps. The trainer's value is no longer programming expertise but accountability, presence, and the social dynamic that drives adherence over months and years.
Extracted from The AI Daily Brief: Artificial Intelligence News, building on Alex Imas's essay 'What Will Be Scarce' from the University of Chicago Booth School of Business.