L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 2 · Lesson 1

The Task Decomposition Principle

Why jobs survive automation even when most of their tasks don't
What is the real unit of automation — and why does the answer change everything?

When Citibank installed New York City's first ATM network in 1977, most economists predicted bank teller employment would collapse within a decade. The machine handled the core task — cash dispensing — faster, cheaper, and without sick days. By 2010, the United States had roughly 400,000 ATMs. It also had more bank tellers than it had in 1980.

The explanation, documented by economist James Bessen in his 2015 analysis of Bureau of Labor Statistics occupational data, was task decomposition: ATMs automated the cash-handling task, which reduced the cost of running a branch, which allowed banks to open more branches, which increased demand for tellers who now spent more time on relationship tasks — loan counseling, account troubleshooting, upselling — that the machines could not do.

The Foundational Distinction

Economists David Autor, Frank Levy, and Richard Murnane introduced the formal task-based framework in their landmark 2003 paper "The Skill Content of Recent Technological Change" published in the Quarterly Journal of Economics. Their core claim: technology automates tasks, not jobs. A job is a bundle of tasks. Automation changes which tasks are inside the bundle and which tasks machines handle — it rarely eliminates the bundle entirely.

This distinction seems obvious once stated, but it has profound implications. When a radiologist's image-reading task is partially automated by AI (as demonstrated by Rajpurkar et al. at Stanford in 2017 with their CheXNet pneumonia detection model), the radiologist's job does not disappear. Instead, the task composition of the role shifts: less time scanning, more time on ambiguous cases, patient communication, and clinical integration. The bundle restructures.

Key Framework — Autor, Levy & Murnane (2003)

Tasks fall into four categories: routine cognitive (rule-based mental work), routine manual (rule-based physical work), non-routine cognitive (judgment, creativity, complex analysis), and non-routine manual (physical tasks requiring situational adaptation). The framework predicts that automation pressure concentrates on routine tasks in both cognitive and manual domains — a prediction confirmed across four decades of U.S. labor data.

Task Exposure vs. Job Exposure

The 2023 GPT-4 paper from OpenAI and University of Pennsylvania researchers (Eloundou, Manning, Mishkin, and Rock) introduced the concept of exposure — distinguishing between tasks that AI can do and tasks that AI can do at lower cost than humans. They found that approximately 80% of U.S. workers have at least 10% of their job tasks exposed to GPT-4 capabilities. But "exposed" does not mean "eliminated." A task being automatable and a job being automated are separated by economics, institutional inertia, regulatory constraints, and the restructuring dynamic the ATM example illustrates.

Similarly, a 2023 Goldman Sachs research report estimated that AI could automate the equivalent of 300 million full-time jobs worth of tasks globally — but explicitly framed this as task exposure, noting that most affected workers would shift into other tasks within their roles or into roles that currently don't exist at scale.

Why This Matters Right Now

Between 2022 and 2024, enterprises began deploying large language models for tasks including document summarization, first-draft contract generation, customer email triage, and code completion. In each case, the pattern matched the task decomposition prediction: paralegal firms using Harvey AI (launched 2022) reported that junior lawyers spent less time on document review and more time on strategic client work. GitHub Copilot users at Microsoft (2023 internal study) wrote code faster but spent more time reviewing, debugging, and designing — tasks the model couldn't reliably handle.

The critical analytical skill — for workers, managers, and policymakers — is learning to decompose jobs into their constituent tasks before predicting automation impact. A job title tells you very little. The task bundle tells you almost everything.

Task exposure:
The degree to which a specific task can be performed by AI at equivalent or lower cost than a human — distinct from job elimination.
Task complementarity:
The phenomenon where automating one task increases demand for human performance of related non-automated tasks in the same role.
Bundle restructuring:
The process by which a job's task composition changes in response to automation without the job itself disappearing.
Real Case — Legal Research

Casetext launched CoCounsel in March 2023, an AI tool that passed the bar exam and could perform legal research tasks in minutes that previously took associates hours. By Q4 2023, the company reported law firm customers were not reducing headcount — they were using the time savings to take on more cases per attorney. Task automation increased attorney throughput; it did not replace attorneys. Thomson Reuters acquired Casetext for $650 million in August 2023, specifically for this task-augmentation model.

Lesson 1 Quiz

The Task Decomposition Principle — four questions
1. According to James Bessen's analysis of ATM deployment, what happened to U.S. bank teller employment between 1980 and 2010?
Correct. This is the canonical example of task complementarity — automating cash dispensing made branches cheaper to run, expanding the branch network and increasing demand for tellers doing non-automated relational tasks.
Not quite. Bessen's data showed teller employment actually grew. ATMs automated the cash-dispensing task but increased demand for tellers by reducing the cost of operating branches. The bundle restructured rather than collapsed.
2. The Autor, Levy & Murnane (2003) framework predicts that automation pressure concentrates most heavily on which category of tasks?
Correct. The ALM framework specifically identifies routine cognitive and routine manual tasks as most automatable, because they can be reduced to explicit rules that machines can follow. Non-routine tasks require judgment, adaptation, or creativity that resists codification.
Reconsider. The ALM framework draws a sharp line between routine and non-routine tasks. Routine tasks — whether cognitive or physical — are most vulnerable because they follow rules that can be encoded. Non-routine tasks, especially those requiring judgment or physical adaptation, are much more resistant.
3. The 2023 Eloundou et al. paper found that approximately what share of U.S. workers have at least 10% of their job tasks exposed to GPT-4 capabilities?
Correct. The paper found roughly 80% exposure at the 10% threshold — a striking figure, but the authors were careful to distinguish task exposure from job elimination, emphasizing that automatable tasks do not automatically translate into displaced workers.
The correct figure from the Eloundou et al. (2023) paper is approximately 80%. This high number is why the authors were careful to emphasize that task exposure and job automation are not the same thing — economic, institutional, and structural factors separate the two.
4. What does "task complementarity" mean in the context of automation economics?
Correct. Task complementarity is the mechanism behind the ATM-teller paradox. When cash dispensing was automated, demand rose for the relational and advisory tasks that tellers do — tasks that complemented, rather than competed with, machine capabilities.
Not quite. Task complementarity describes a dynamic where automating Task A increases demand for Task B because both belong to the same role and the efficiency gain from automating A makes the overall role more valuable or more widely deployed. The ATM-teller example is the classic illustration.

Lab 1 — Task Decomposition Practice

Break a job into its task bundle. Identify what's routine vs. non-routine.

Your Mission

Choose any job title — your own, a colleague's, or one you're curious about. Work with the AI assistant to decompose it into individual tasks, then classify each task using the ALM framework (routine cognitive, routine manual, non-routine cognitive, non-routine manual). Finally, assess which tasks have the highest AI exposure today.

Try: "Help me decompose the job of [your title] into its component tasks, then classify each using the Autor-Levy-Murnane framework." After that, ask which tasks are most exposed to current AI tools and why.
Task Decomposition Assistant
ALM Framework
Hello! I'm here to help you decompose any job into its task bundle using the Autor-Levy-Murnane framework. Tell me a job title — yours, a colleague's, or any role you're curious about — and we'll break it down together. Which tasks are routine? Which are non-routine? Which face the highest AI exposure? Let's find out.
Module 2 · Lesson 2

The Middle-Skills Hollowing

Why automation hits the middle of the wage distribution hardest
If automation is task-based, which jobs are actually disappearing — and why are they clustered where they are?

Between 2000 and 2015, U.S. manufacturing employment fell from approximately 17.3 million to 12.3 million workers — a loss of five million jobs. MIT economist David Autor's research, published in the Journal of Economic Perspectives in 2015, demonstrated that this collapse was not uniform across the wage spectrum. The losses were concentrated almost entirely in middle-wage, middle-skill occupations: machine operators, assembly-line workers, clerical processors, data entry clerks — all roles dense with routine tasks.

High-skill professional jobs grew. Low-skill service jobs grew. The middle contracted sharply. Economists named this pattern labor market polarization — and the mechanism was task automation targeting the routine-task-intensive jobs that happened to cluster in the middle of the wage distribution.

Polarization: The Evidence

Autor and Dorn's 2013 paper "The Growth of Low-Skill Service Jobs and the Polarization of the U.S. Labor Market" in the American Economic Review established the statistical pattern rigorously. Between 1980 and 2005, employment in occupations paying middle wages (roughly the 30th–70th percentile) shrank as a share of total employment, while both the top and bottom grew. The jobs that disappeared were not random — they were specifically jobs with high routine-task content.

The pattern has been replicated across European labor markets. Goos, Manning, and Salomons (2009) found similar polarization in 16 European countries between 1993 and 2006. The consistency across different labor market institutions — different minimum wages, union densities, social protections — suggests the mechanism is fundamentally technological, not regulatory.

5M
U.S. Mfg Jobs Lost
2000–2015, concentrated in routine-task roles
16
European Countries
Showing identical polarization pattern (Goos et al. 2009)
47%
U.S. Jobs at High Risk
Frey & Osborne 2013 — routine-task concentration metric
The Frey & Osborne Methodology (2013)

Oxford researchers Carl Benedikt Frey and Michael Osborne published "The Future of Employment" in September 2013. They rated 702 U.S. occupations by computerization susceptibility, finding 47% of jobs at high risk. Critics noted the study assessed tasks within jobs imprecisely — but it sparked the research wave that led to more rigorous task-level analyses. Its core insight — that routine-task density predicts automation risk — remains the foundation of subsequent work.

Why the Middle? The Wage-Routine Correlation

The reason automation targets the middle of the wage distribution is a structural feature of how labor markets price tasks. Routine tasks command middle wages because they require specific training but not advanced judgment — enough skill to be valued, not enough complexity to command a premium. Non-routine cognitive tasks (analysis, management, design) command high wages. Non-routine manual tasks (building maintenance, elder care, food service) command low wages because they require physical presence and situational adaptation but not credentialed expertise.

When computers and robots automate routine tasks, they eliminate the middle layer. This creates a barbell-shaped labor market: growing demand at the high-skill end and growing (low-wage) demand at the low-skill end, with the middle contracting. For workers whose skills were priced in the middle band, the adjustment is brutal — either significant upskilling toward the high end or downward movement to lower-paying service work.

The New Wave: Cognitive-Middle Exposure

The first wave of automation (1980–2015) hit routine manual and routine cognitive work. The AI wave of 2020–present extends the threat upward into what was previously "safe" non-routine cognitive territory. Tasks like first-draft legal writing, basic financial analysis, standard radiology reads, and entry-level code generation are now automatable in ways they were not under prior technologies.

A 2023 study by Acemoglu and Restrepo estimated that AI exposure is now strongest in occupations at the 60th–80th percentile of the wage distribution — professional and semi-professional workers who historically assumed automation would not reach them. The hollowing effect is migrating upward. This does not mean wholesale job elimination, but it does mean significant task-bundle restructuring in white-collar work that mirrors what happened to blue-collar work a generation earlier.

Occupational Group Wage Tier 1980–2015 Impact AI-Era Exposure
Machine operators, assembly Middle High — major job loss Continues, robotics
Clerical, data entry, bookkeeping Middle-low High — major job loss Very high — LLMs
Lawyers, doctors, financial analysts High Low — minimal impact Moderate — task exposure
Food service, elder care, cleaning Low Low — limited automation Low — physical adaptation
Paralegals, junior accountants Middle-high Low historically High — LLM document tasks
Key Insight — Accenture / WEF Research

The World Economic Forum's "Future of Jobs Report 2023" surveyed 803 companies employing 11.3 million workers across 45 economies. It projected that 26% of current tasks would be automated by 2027, with the heaviest impact on information processing and clerical tasks — confirming the polarization prediction extends into the current AI cycle. The same report predicted 69 million new jobs created by 2027, primarily in green energy, data, and AI development — but these require substantially different skills than the roles being displaced.

Lesson 2 Quiz

The Middle-Skills Hollowing — four questions
1. "Labor market polarization" describes which specific pattern in employment data?
Correct. Polarization means a barbell shape: the high-skill and low-skill ends grow while the middle contracts. Autor and Dorn (2013) documented this rigorously across U.S. data from 1980–2005, and Goos et al. (2009) replicated the finding across 16 European countries.
Not quite. Polarization is the opposite of what the first option describes. The middle shrinks — high-wage professional jobs and low-wage service jobs both grow while middle-wage routine-task roles decline. The mechanism is automation targeting routine tasks, which happen to cluster at middle wages.
2. Why do routine tasks tend to cluster in the middle of the wage distribution?
Correct. The wage-routine correlation reflects labor market pricing: routine tasks require enough skill to be compensated above minimum wage but lack the complexity, creativity, or judgment that commands professional-level pay. This structural feature is why automation hits the middle hardest.
The core explanation is about task characteristics and market pricing, not institutional factors. Routine tasks sit in the middle of the skill spectrum — trained but codifiable — so labor markets price them in the middle wage band. When those tasks are automatable, middle-wage jobs disappear.
3. Frey and Osborne's 2013 Oxford study found what percentage of U.S. jobs at high computerization risk?
Correct. Frey and Osborne rated 702 occupations and found 47% at high computerization risk. The study was influential but also criticized for imprecise task-level assessment. Its lasting contribution was establishing routine-task density as the key predictor of automation risk.
The Frey and Osborne (2013) figure was 47%. This number sparked considerable debate — critics argued the occupation-level analysis was too coarse and overstated risk compared to task-level analyses — but the core finding that routine-task density predicts susceptibility has held up in subsequent research.
4. According to Acemoglu and Restrepo (2023), the current AI wave is creating the heaviest task exposure at which part of the wage distribution?
Correct. The AI wave is migrating the automation frontier upward into the professional middle — paralegals, junior analysts, entry-level coders, radiologists for standard reads. This mirrors what automation did to blue-collar routine work in the 1980s–2000s, but now applied to white-collar work.
Acemoglu and Restrepo found the current AI wave concentrates exposure in the 60th–80th wage percentile — professional and semi-professional workers who historically assumed they were safe from automation. The hollowing effect is moving up the wage ladder, which is why this wave feels different to many white-collar workers.

Lab 2 — Polarization Mapping

Analyze how automation has reshaped a specific industry's job mix over time.

Your Mission

Pick an industry you know or care about — banking, healthcare, legal services, retail, manufacturing, or any other. Work with the AI to map how that industry's job composition has changed due to automation over the past 20 years, and where the current AI wave is applying new pressure. Look specifically for the polarization signature: which middle-skill roles have shrunk, and what grew at the top and bottom?

Try: "Walk me through how automation has polarized the job mix in [industry] between 2000 and today. Which specific roles disappeared? What grew? And where is AI applying pressure now?"
Polarization Analysis Assistant
Labor Market Data
Ready to map polarization in any industry you choose. Tell me which sector interests you — banking, healthcare, legal, retail, manufacturing, logistics, media — and I'll walk through how automation reshaped its job mix from 2000 to today, identifying which middle-skill roles contracted, what grew, and where the current AI wave is hitting hardest.
Module 2 · Lesson 3

When Automation Creates Jobs

The new-task mechanism and the history of technology-driven employment growth
Has technology ever created more jobs than it destroyed — and what conditions make that happen?

VisiCalc, the first commercial spreadsheet software, launched in 1979. By eliminating the manual calculation tasks that consumed most of a bookkeeper's day, economists predicted a collapse in accounting employment. Instead, the number of accounting professionals in the United States more than doubled between 1980 and 2010, from roughly 1.3 million to over 2.7 million.

The mechanism: spreadsheets so dramatically reduced the cost of financial analysis that demand for financial analysis exploded. Tasks that were previously too expensive to perform — detailed variance analysis, scenario modeling, monthly reporting for small businesses — became economically viable. New tasks created demand for more accountants doing more sophisticated work than their predecessors ever performed.

Acemoglu and Restrepo: The New-Task Framework

MIT economists Daron Acemoglu and Pascual Restrepo published "The Race Between Man and Machine" in the American Economic Review in 2018, introducing a formal model for understanding when automation creates versus destroys jobs net. Their core insight: technology operates through two simultaneous forces — a displacement effect (automating existing tasks reduces demand for human labor in those tasks) and a reinstatement effect (new technologies create new tasks that require human labor).

Whether net employment rises or falls depends on which effect dominates. In the spreadsheet case, the reinstatement effect was enormous because cheap computation unlocked an entire new category of analytical work. In U.S. manufacturing from 2000–2015, the displacement effect dominated because new tasks (software development, data science) required completely different skills and geography than the jobs being eliminated.

The New-Task Mechanism — Documented Examples

E-commerce and logistics (2010–2023): Amazon's warehouse robotics (Kiva Systems, acquired 2012) automated picking-path optimization. Simultaneously, e-commerce growth created enormous demand for delivery drivers, last-mile logistics coordinators, and returns management specialists — jobs that didn't exist at scale before 2005. U.S. courier and delivery employment grew from 525,000 in 2010 to over 1.2 million by 2022 (BLS data).

Social media moderation (2010–present): AI content classification tools handle the majority of spam and bot detection. This created an entirely new occupation — human content moderator — that now employs an estimated 100,000+ people globally (NYU Stern Center for Business and Human Rights, 2021 report) to handle cases that AI cannot reliably classify.

The Historical Record on Technology and Employment

Economic historian Robert Gordon's monumental 2016 work The Rise and Fall of American Growth documents the employment effects of the major technological waves since 1870. The pattern is consistent: each wave — electrification, mass production, computing — initially displaces existing workers in specific roles and eventually creates new categories of work that expand total employment. The key variable is the transition period: how long it takes for new tasks to absorb displaced workers, and whether those workers have the skills and geographic access to perform the new tasks.

The economist Erik Brynjolfsson and his co-authors have documented what they call the productivity paradox: major technologies often show weak productivity statistics for years after deployment because organizations haven't yet restructured around them to create the new tasks. The steam engine's productivity gains were negligible for the first 30 years; general-purpose computing showed no clear productivity dividend until the mid-1990s, roughly 30 years after the first mainframes. AI productivity gains are currently in this early adoption phase.

New Jobs Created by AI — Current Evidence

Since 2020, several new occupational categories have emerged at scale that did not meaningfully exist before. Prompt engineers — specialists in designing inputs to maximize AI output quality — were categorized as a distinct role by LinkedIn in 2023 and showed 1,400% job posting growth between Q1 2022 and Q1 2023. AI trainers and AI output evaluators (roles at companies like Scale AI, Surge AI, and Anthropic's RLHF programs) employ tens of thousands of workers to provide human feedback that improves model quality.

AI compliance officers and algorithmic auditors have emerged in response to the EU AI Act (passed 2024) and similar regulations globally. Consulting firm Gartner projected in 2023 that AI regulatory compliance would become a dedicated function in 30% of large enterprises by 2025. These roles require domain expertise plus AI understanding — a combination that currently commands significant wage premiums.

Key Tension — Speed of Reinstatement

The critical unresolved question is speed. The spreadsheet took 15 years to create its full employment expansion. E-commerce logistics took roughly 10 years. AI's displacement effects are measurable now; its reinstatement effects are still forming. Acemoglu's 2022 update to his framework noted that if AI primarily automates existing tasks without creating genuinely new task categories at comparable scale, the net employment effect could be negative in ways previous automation waves were not. This remains an active debate among labor economists — not settled science.

Displacement effect:
Automation reduces demand for human labor in tasks that machines now perform — a direct job-reduction force.
Reinstatement effect:
New technologies create new tasks that did not previously exist, generating fresh demand for human labor and potentially offsetting displacement.
Productivity paradox:
Major general-purpose technologies often show weak productivity statistics for decades after introduction, before organizational restructuring unlocks their full gains.

Lesson 3 Quiz

When Automation Creates Jobs — four questions
1. In the Acemoglu-Restrepo (2018) framework, what is the "reinstatement effect"?
Correct. The reinstatement effect is the positive force that can offset displacement — new technologies create categories of work that didn't exist before, opening demand for human labor. Whether net employment rises or falls depends on whether reinstatement outpaces displacement in a given technology wave.
In Acemoglu and Restrepo's framework, the reinstatement effect specifically means new technologies generating new tasks — categories of work that simply didn't exist before. This is what happened with accounting and spreadsheets: cheap computation created entirely new financial analysis tasks that accountants hadn't previously performed.
2. What happened to U.S. accounting employment between 1980 and 2010, following the introduction of spreadsheet software in 1979?
Correct. The spreadsheet more than doubled accounting employment because it reduced the cost of financial analysis so dramatically that demand for analysis expanded enormously — making tasks economically viable that were previously too expensive. This is the reinstatement effect at work.
The counterintuitive answer is that accounting employment more than doubled. Spreadsheets reduced cost so sharply that the quantity demanded of financial analysis exploded — small businesses could now afford analysis that only large firms previously justified, and entirely new categories of analytical work became economically viable.
3. The "productivity paradox" documented by Brynjolfsson refers to which phenomenon?
Correct. The paradox is the long lag between technology deployment and measurable productivity gains. Steam engines showed minimal productivity impact for 30 years; general-purpose computers showed no clear dividend until the mid-1990s, decades after their introduction. Organizations need time to restructure around new technologies before gains appear in data.
The productivity paradox refers to the historical pattern where major technologies — steam, electricity, computing — show surprisingly weak productivity statistics for many years or even decades after widespread deployment. Brynjolfsson documented this for computing in the 1980s–90s. The explanation is that organizational restructuring, which is what actually delivers gains, takes time.
4. U.S. courier and delivery employment grew from 525,000 in 2010 to over 1.2 million by 2022 — this is an example of which mechanism?
Correct. Last-mile delivery is a reinstatement-effect example. Amazon's warehouse robotics automated picking paths, but the e-commerce volume those efficiencies enabled created an entirely new category of work — home delivery at scale — that didn't exist meaningfully in 2005. Technology created the task that employs the workers.
This is a reinstatement-effect example. Amazon warehouse automation reduced labor intensity per package, but the cost reductions and scale gains enabled e-commerce to grow enormously, creating massive demand for a new task — last-mile home delivery — that barely existed before 2010. New technology created new tasks, which created jobs.

Lab 3 — Displacement vs. Reinstatement Analysis

Map the full economic impact of a specific AI deployment: what's displaced, what's reinstated?

Your Mission

Choose a specific AI tool or deployment — GitHub Copilot, Harvey AI, GPT-4 in customer service, radiological AI, or any other. Work with the AI to map both the displacement effect (which tasks and roles face reduced demand) and the reinstatement effect (which new tasks or roles are being created or expanded). Then assess whether the net effect is likely positive, negative, or ambiguous based on current evidence.

Try: "Analyze [specific AI tool] using the Acemoglu-Restrepo framework. What tasks does it displace? What new tasks or roles is it creating? Is the net employment effect positive, negative, or too early to call — and why?"
Displacement-Reinstatement Analyst
Acemoglu-Restrepo Model
Let's apply the Acemoglu-Restrepo displacement-versus-reinstatement framework to a real AI deployment. Name any specific tool — GitHub Copilot, Harvey AI, Salesforce Einstein, ChatGPT in customer service, warehouse robotics, radiology AI — and I'll map what it displaces, what new tasks or roles it creates, and whether the net effect looks positive, negative, or genuinely ambiguous given current evidence.
Module 2 · Lesson 4

Reading Your Own Exposure

How to assess your personal task bundle and adapt strategically
Given everything we know about how automation actually works, what should you actually do about your own career?

In November 2016, AI pioneer Geoffrey Hinton told the BBC: "We should stop training radiologists now. It is quite obvious that within five years deep learning is going to do better than radiologists." By 2024, radiology had one of the worst physician shortages in U.S. medicine — an estimated shortage of 42,000 radiologists projected by the American College of Radiology for 2030.

What Hinton missed: AI improved radiology read accuracy on specific image types — CheXNet (2017) at pneumonia detection, Paige AI at prostate cancer detection (FDA-cleared 2021) — but did not automate the full task bundle. Radiologists increasingly spend time on clinical integration, complex multi-organ reads, interventional procedures, peer consultation, and AI output validation. The task bundle restructured. Demand for radiologists grew precisely because AI-assisted reads became so fast that volumes increased, creating more work for human physicians.

A Personal Exposure Assessment Framework

The research provides a practical framework for assessing your own exposure. The Eloundou et al. (2023) methodology can be adapted personally: for each task in your job, ask three questions.

1. Is this task codifiable? Can the task be fully described by explicit rules, patterns, or examples that an AI could learn from? Data entry, standard form completion, and template drafting score high. Negotiating a complex deal, diagnosing an ambiguous situation, or building trust with a client score low.

2. Does this task require presence or embodiment? Physical presence, tactile judgment, real-time situational awareness, and direct human connection are difficult for AI to replicate. Tasks requiring these qualities maintain human advantage regardless of software sophistication.

3. Is this task the bottleneck? Even if a task is automatable, if it's the rate-limiting step in a process dominated by non-automatable tasks, its automation may create task complementarity rather than displacement — speeding up the automatable portion to free time for the parts that can't be automated.

Case Study — Paralegal Work at Allen & Overy

In November 2023, global law firm Allen & Overy announced a partnership with Harvey AI to deploy the tool across its 3,500+ lawyers. The firm explicitly stated it would not reduce headcount. Instead, it tracked how attorneys used freed time. Six months later, internal data showed partners were using Harvey for first-draft contract generation and standard clause research, spending that recovered time on client relationship management, regulatory strategy, and cross-practice coordination — tasks the model could not perform. The task bundle shifted. Billable hours stayed flat or grew slightly; the composition of what filled those hours changed substantially.

The Skills That Complement AI — Evidence-Based

The MIT Work of the Future Task Force's 2023 annual report, based on surveys of 15,000 workers across 10 countries, identified the skill clusters that showed the strongest wage growth alongside AI adoption. Three patterns emerged consistently.

Contextual judgment in AI-augmented workflows. Workers who can correctly identify when AI output is wrong — catching model errors, hallucinations, or context mismatches — showed significantly higher productivity and retention than those who simply used AI output uncritically. This requires deep domain knowledge, not just AI familiarity.

Complex coordination and orchestration. As individual tasks become cheaper, coordinating the overall process across stakeholders, managing exceptions, and integrating AI outputs into decisions grows in relative importance. Project management, account management, and cross-functional coordination roles showed strong demand growth in every country surveyed.

Relationship-intensive tasks with accountability requirements. Tasks where trust, accountability, and relationship continuity matter — client-facing work, management, negotiation, ethics oversight — proved highly resistant to AI substitution. These tasks require not just competence but the human credibility and accountability that institutions and clients demand.

What the Research Does Not Say

It's worth being clear about what the automation research cannot tell you. It cannot predict your specific career trajectory. Labor economics describes aggregate patterns across large populations; individual outcomes depend on firm-level decisions, regional labor markets, personal networks, and policy choices that aggregate models cannot capture. The same task bundle faces very different outcomes depending on whether your employer views AI as a cost-reduction tool or a productivity-expansion tool.

The research also does not support either extreme interpretation — neither "AI will eliminate most jobs" nor "AI creates only opportunity." The honest summary from the evidence: significant task-level restructuring is certain; net job destruction at aggregate scale is not established by current data but remains plausible under some AI development trajectories; the key variable is how quickly reinstatement-effect job creation scales relative to displacement-effect job reduction. Workers who understand this framework are positioned to navigate it more strategically than those operating on hype in either direction.

Practical Implication — The Portfolio Approach

Based on the task decomposition and polarization research, career resilience in an AI-augmented labor market involves maintaining a portfolio of tasks rather than a single skill. Specifically: one anchor of deep domain expertise that gives you the contextual judgment to evaluate AI outputs; one relationship-intensive competency that is structurally resistant to automation; and active AI tool fluency that lets you leverage automation to expand your throughput rather than compete with it. This portfolio approach is consistent with what the MIT Work of the Future data shows about the workers who are adapting most successfully.

Lesson 4 Quiz

Reading Your Own Exposure — four questions
1. By 2024, what had actually happened to U.S. radiology employment following Hinton's 2016 prediction that AI would replace radiologists within five years?
Correct. Radiology illustrates task-bundle restructuring rather than job elimination. AI improved accuracy on specific image types but couldn't automate the full bundle — and the efficiency gains increased imaging volumes, expanding total demand. The result was an acute shortage, not surplus, of radiologists.
The outcome was the opposite of Hinton's prediction: severe shortage. AI automated specific narrow tasks (pneumonia detection, prostate cancer reads) but couldn't handle the full radiology task bundle. Faster AI-assisted reads enabled more imaging volume, increasing total work. The ACR projected a 42,000 physician shortage by 2030.
2. In the personal exposure assessment framework, a task that requires real-time situational awareness, physical presence, and tactile judgment would score how on automation risk?
Correct. Tasks requiring physical presence, tactile judgment, and real-time situational adaptation remain highly resistant to automation. This is a structural feature, not a temporary gap — the combination of physical embodiment, sensorimotor integration, and situational adaptation is genuinely difficult for current AI and robotic systems.
Physical presence, tactile judgment, and real-time situational adaptation score low on automation risk. These require embodiment and sensorimotor integration that current AI and robotics handle poorly. The ALM framework's "non-routine manual" category — cleaning, elder care, plumbing, surgery — remains resistant to automation precisely because of these characteristics.
3. The MIT Work of the Future Task Force (2023) identified which skill as showing particularly strong wage growth alongside AI adoption?
Correct. Workers who can catch AI errors — hallucinations, context mismatches, domain-specific mistakes — using deep expertise showed significantly higher productivity and retention. This requires knowing the domain well enough to recognize when the AI is confidently wrong, which is a distinctly human skill in AI-augmented work.
The MIT Work of the Future data specifically highlighted contextual judgment — the ability to recognize AI errors using deep domain expertise — as a high-value skill alongside AI adoption. Simply using AI tools is not sufficient; the workers adding most value are those who know their domains well enough to catch and correct AI mistakes.
4. What does the current body of automation research conclusively establish about net job destruction from AI?
Correct. This is the honest, evidence-consistent summary: task restructuring is clear and documented; net aggregate job destruction is neither proven nor disproven. The outcome depends on how quickly the reinstatement effect scales relative to displacement — which in turn depends on AI development trajectories and policy choices that are not predetermined.
Neither extreme is supported by current evidence. The research is clear that task restructuring is certain and significant. But net job destruction at aggregate scale is not established — the outcome depends on whether AI triggers a robust reinstatement effect (new tasks) comparable to previous technology waves. This is a live empirical question, not settled science in either direction.

Lab 4 — Personal Exposure Assessment

Apply the full module framework to assess your own task bundle and build a resilience strategy.

Your Mission

This lab applies everything from the module to your specific situation. Describe your current role or a role you're targeting. The AI assistant will help you decompose it into tasks, classify exposure using the ALM framework, identify which tasks face the highest AI pressure, and develop a concrete strategy for leveraging the module's insights — task complementarity, bundle restructuring, the portfolio approach — for your own career.

Start with: "I work as [your role] and my main responsibilities include [brief description]. Help me assess my task-level AI exposure and build a resilience strategy based on what the automation research actually says."
Personal Exposure Strategist
Career Analysis
This is the practical capstone of the module. Tell me about your role — your title, your main responsibilities, the tasks you spend most time on. I'll help you decompose your job into its task bundle, assess each task's AI exposure using the ALM framework, identify where you face genuine risk vs. where you have structural protection, and develop a concrete strategy using what the research actually shows about task complementarity, bundle restructuring, and the portfolio approach. What's your role?

Module 2 Test

Task Automation vs. Job Automation — 15 questions · 80% to pass
1. The core claim of the Autor, Levy & Murnane (2003) framework is best summarized as:
Correct.
The ALM core claim: technology automates tasks, and jobs are task bundles. Automation changes which tasks are in the bundle, rarely eliminating bundles entirely.
2. "Task exposure," as used in the Eloundou et al. (2023) paper, most precisely means:
Correct.
Task exposure is the degree to which AI can perform a task at equivalent or lower cost — distinct from job elimination. High exposure doesn't mean the job disappears; it means that task may shift to AI within a restructured bundle.
3. James Bessen's analysis of ATM deployment demonstrates which principle?
Correct.
The ATM case demonstrates task complementarity: automating cash dispensing reduced branch costs, enabled branch expansion, and increased demand for tellers who now performed relationship tasks the machines couldn't do.
4. Labor market polarization predicts employment growth in which two areas simultaneously?
Correct.
Polarization means a barbell: both high-skill and low-skill employment grow while the middle (routine-task-intensive) contracts. This creates a hollowed labor market, not a simple high-to-low shift.
5. Goos, Manning, and Salomons (2009) found labor market polarization in how many European countries?
Correct.
Goos et al. documented polarization across 16 European countries with very different labor market structures, suggesting the mechanism is primarily technological rather than regulatory or institutional.
6. Why do routine tasks cluster specifically in the middle of the wage distribution?
Correct.
The wage-routine correlation reflects task characteristics: enough training to command middle wages, but codifiable enough that machines can eventually replace them — which is precisely why automation hits the middle hardest.
7. The Acemoglu-Restrepo (2018) "Race Between Man and Machine" framework centers on the balance between which two forces?
Correct.
Acemoglu and Restrepo's framework identifies displacement (automation reduces human labor demand in automated tasks) and reinstatement (new tasks create new human labor demand) as the competing forces determining net employment outcomes.
8. U.S. accounting employment roughly doubled between 1980 and 2010 despite the introduction of spreadsheet software because:
Correct.
The reinstatement effect: spreadsheets made analysis cheap enough that businesses that previously couldn't justify financial analysis now could. New tasks (scenario modeling, monthly variance analysis for small firms) were created by the technology, expanding demand for the profession.
9. Erik Brynjolfsson's "productivity paradox" refers to:
Correct.