The “humanity discount” and the quiet repricing of knowledge work
Svetlana Makarova, an AI product leader at IKS Health, is putting a precise label on a diffuse workplace sensation: the growing “humanity discount.” The idea is deceptively simple—once human output is routinely compared to machine consistency, the very traits that make people valuable in complex environments (variability, judgment calls, emotional nuance) can be reframed as defects rather than strengths.
In practice, this discount shows up as a subtle repricing of human capital. Organizations that once rewarded adaptability and intuition may increasingly reward predictability, repeatability, and compliance with standardized workflows—not because leaders explicitly devalue people, but because AI systems make variance more visible and easier to penalize. When a model can draft, summarize, classify, or triage at scale, the baseline expectation for throughput rises. The human worker becomes the “exception handler,” yet is often evaluated with metrics designed for the “happy path.”
This is not merely a cultural shift; it is a management shift. Talent systems built around benchmarking, performance dashboards, and output quotas can unintentionally import machine logic into human evaluation. The risk is that firms begin to treat human contribution as a set of modular tasks—useful only insofar as it resembles what an algorithm can already do reliably.
Standardization didn’t just enable AI—it trained it
Makarova’s framing also highlights an uncomfortable continuity: today’s AI acceleration is less a sudden disruption than the culmination of decades of workflow codification. Many industries have spent years decomposing knowledge work into measurable units:
- Sales: scripted outreach, standardized qualification criteria, CRM-driven funnel metrics
- Legal: templates, clause libraries, document review protocols, playbooks
- R&D and product: ticketing systems, structured experimentation logs, KPI-driven roadmaps
- Operations and support: runbooks, escalation matrices, QA rubrics, time-to-resolution targets
These practices were adopted for efficiency, quality control, and scalability. Yet they also produced something else: structured data and repeatable patterns—the raw material modern AI systems learn from and operate within. In effect, organizations have been building the conditions for automation for years, often without explicitly intending to.
The strategic implication is that AI adoption is not simply about buying tools; it is about recognizing that many enterprises have already reorganized work into automation-ready components. That makes deployment faster—but it also raises the stakes. When tasks are already standardized, AI doesn’t just assist; it can substitute. And when substitution becomes feasible, the bargaining power of routine knowledge work can weaken, even if headcount reductions are not immediate.
Emotional asymmetry: when machines “perform empathy” and humans absorb the cost
A particularly sharp edge of the humanity discount emerges in the emotional layer of work. As AI agents become more conversational and “supportive” in tone, workplaces may drift into a new kind of asymmetry: humans are expected to provide emotional intelligence, resilience, and interpersonal finesse—while receiving little genuine reciprocity in return.
This matters because emotional labor is not an abstract concept in modern organizations; it is embedded in daily execution:
- managing stakeholder expectations amid ambiguity
- de-escalating conflict and maintaining trust
- mentoring, coaching, and onboarding
- navigating organizational politics and cross-functional tradeoffs
AI can simulate empathy through language, but it does not share the human burden of accountability, fatigue, or reputational risk. If leaders treat AI’s “pleasant interface” as a substitute for human support systems, the result can be a workplace that feels efficient yet psychologically extractive—high output, low reciprocity.
Makarova’s warning intersects with a broader operational reality: when AI increases throughput, organizations often respond by raising targets. The promise that automation will “free people for higher-value work” can morph into an efficiency treadmill, where time saved is immediately reallocated to more tasks, more monitoring, and tighter deadlines. Over time, this can compress the space required for deep thinking, creativity, and skill development—the very capacities companies claim they want more of.
What organizations and professionals can do before the discount becomes doctrine
The most actionable element in Makarova’s analysis is the implied choice: the humanity discount is not inevitable, but it is highly compatible with how many firms already manage performance. Avoiding it requires intentional design—of roles, metrics, and governance—so that AI augments human capability rather than redefining humans as “less reliable machines.”
For organizations, several moves stand out:
- Redesign roles around human advantage, not just automated task removal: strategic judgment, cross-functional synthesis, ethical oversight, and decision-making under uncertainty.
- Build metric sanity checks into performance management by blending quantitative KPIs with qualitative signals such as peer review, customer feedback, and innovation audits.
- Treat emotionally engaging AI as a workforce intervention that requires safeguards: transparency, escalation paths, feedback loops, and well-being instrumentation that detects burnout rather than merely measuring output.
- Engage early with policy and regulation on algorithmic accountability, data rights, and reskilling incentives, shaping rules that protect trust without freezing innovation.
- Align AI gains with social sustainability through inclusive upskilling and equitable value-sharing, especially in environments where labor’s share of productivity gains is already under pressure.
For professionals navigating AI-driven workplaces, Makarova’s prescription is pragmatic: double down on what remains hard to automate. That includes deep domain expertise, social acuity, and the ability to make decisions when the data is incomplete, the incentives conflict, and the outcome is uncertain. In a world optimized for machine consistency, the differentiator is not being more machine-like—it is being reliably human in the moments that matter.
The companies that will lead the next phase of digital transformation are unlikely to be those that merely deploy AI fastest. They will be the ones that resist the temptation to discount humanity—and instead build operating models where human judgment, trust, and accountability are treated not as inefficiencies, but as strategic assets that machines cannot replace.
By
By
By
By
By
By
