Understanding The Economic Trade-Offs Of Automation

As AI is replacing early-career jobs, the economy's productivity in the short-run increases, but productivity and welfare in the long run may decline. In a new paper, Enrique Ide argues that we may be witnessing "socially excessive automation of early-career work. Such automation may deliver immediate productivity gains, but it also erodes the skills of future cohorts and constrains long-run growth." Here's the abstract of his paper: "Recent advances in Artificial Intelligence (AI) have sparked expectations of unprecedented economic growth. Yet, by enabling senior workers to accomplish more tasks independently, AI may reduce entry-level opportunities, raising concerns about how future generations will acquire expertise. This paper develops a model to examine how automation and AI affect the intergenerational transmission of tacit knowledge—practical, hard-to-codify skills critical to workplace success. I show that the competitive equilibrium features socially excessive automation of early-career tasks, and that improvements in such automation generate an intergenerational trade-off: they raise short-run productivity but weaken the skills of future generations, slowing long-run growth—sometimes enough to reduce welfare. Back-of-the-envelope calculations suggest that AI-driven entry-level automation could reduce the long-run annual growth rate of U.S. per-capita output by 0.05 to 0.35 percentage points, depending on its scale. I further show that AI co-pilots can partially offset lost learning by assisting individuals who fail to acquire skills early in their careers. However, they may also weaken juniors’ incentives to develop such skills. These findings highlight the importance of preserving and expanding early-career learning opportunities to fully realize AI’s potential." What can policy do? In the concluding remarks, the paper offers several ideas: - government subsidies for "mentorship, apprenticeship, and other entry-level training arrangements" - "taxing entry-level automation" - reducing minimum wages for young workers - promoting AI systems that complement rather than replace entry-level jobs. Universities could also play a role by placing "greater emphasis on providing undergraduate students with opportunities to gain practical experience before they formally enter the labor market. Such initiatives would complement the traditional focus of undergraduate programs on codifiable knowledge and help foster the early development of tacit skills." Read the full paper here: https://lnkd.in/eMq3uktX (open access)

A new paper from David Autor, in collaboration with Neil Thompson, makes an important contribution to explaining how AI is likely to impact labor markets. Based on a rigorous model, confirmed with an analysis of 40 years of data, they provide a nuanced perspective on how automation impacts job employment and wages. Essentially, this depends on the extent to which easy tasks are removed from a role and expert ones are added, and how specialized a role becomes as a result. When jobs gain inexpert tasks but lose expertise, wages decline, but employment may increase. Think of how taxi driving became less specialized, and well-paid, but more common, due to Uber. In contrast, when technology automates the easy tasks inside a job, the remaining work becomes more specialized. Employment falls because fewer people now qualify, but the scarcity of expertise drives wages up. This is what seems to be happening with proofreading, which is now less about spell-checking and more about helping people to write, leading to lower job numbers but higher average wages. Their model helps us to understand the impacts of AI on labor markets. For instance, why AI tools can raise wages for senior software engineers, but decrease employment, while simultaneously reducing earnings, and increasing employment, for more entry level software engineering roles.

Most people think automation is a growth story. It’s not that simple. A joint study from University of Pennsylvania and Boston University explores a deceptively simple but dangerous question: What happens when companies automate faster than the economy can adapt? The answer isn’t disruption. It’s self-destruction, just slower, and harder to see. Here’s the mechanism no one talks about: When companies automate, they reduce costs. That’s the obvious win. But they also reduce jobs. And jobs are what create purchasing power. So the same system that makes companies more efficient… quietly removes the very customers those companies depend on. At first, everything looks great: • Margins improve • Productivity spikes • Share prices go up It feels like progress. But underneath, something starts to fracture. Fewer people earning → less spending → weaker demand. And demand is the lifeblood of every business. This creates a dangerous feedback loop: Automation → Job loss → Lower consumption → Revenue pressure → More automation A system optimizing for efficiency… begins to cannibalize itself. This isn’t theory. It’s a structural tension inside modern capitalism. Technology moves exponentially. Society adjusts linearly. That gap is where instability lives. The real risk isn’t that AI takes jobs. It’s that we don’t redesign the system around it. Because if productivity gains don’t translate into: • new forms of income • new types of work • or redistributed value Then we’re not creating wealth. We’re concentrating it and shrinking the base that sustains it. The future won’t be decided by how fast we automate. It will be decided by how intelligently we adapt. That means rethinking: • education • income models • business incentives • and what “work” even means Because a system that optimizes for efficiency alone… Eventually runs out of people who can afford to participate in it. And when that happens, the collapse doesn’t look like a crash. It looks like slow, irreversible stagnation.

This paper examines the macroeconomic implications of AI advancements, focusing on productivity, inequality, and the creation of new tasks. 1️⃣ Using a task-based model, the author analyzes how AI-driven automation and task complementarities could impact the economy over the next decade, providing a balanced view of both benefits and limitations. 2️⃣ AI-driven productivity improvements are modest, with estimates of no more than a 0.66% increase in overall productivity over the next 10 years, and a more conservative estimate of around 0.53%. 3️⃣ AI will likely impact 20% of U.S. labor tasks, but only 4.6% of these tasks are expected to be profitably automated within the next 10 years. 4️⃣ Cost savings from AI-driven automation are estimated at 27% of labor costs in exposed tasks, though this varies by industry and task type. 5️⃣ The projected GDP growth due to AI is estimated to be between 0.93% and 1.16% over the next decade, with an upper bound of 1.56% if investment responses are stronger. 6️⃣ AI’s productivity gains are concentrated in “easy-to-learn” tasks—those with clear, objective outcomes—while “hard-to-learn” tasks, requiring complex, context-dependent decisions, will yield smaller gains. 7️⃣ AI’s wage effects are unlikely to reduce inequality. Even when AI improves the productivity of low-skill workers, it could increase income inequality by widening the gap between capital and labor income. 8️⃣ While demographic inequality may not increase as much as with previous technologies like robotics, AI is still expected to favor capital over labor, with limited overall wage growth. 9️⃣ New tasks created by AI could enhance productivity, but some may have negative social value, such as manipulative algorithms or addictive digital content, potentially reducing overall welfare. ✍🏻 Daron Acemoglu. The Simple Macroeconomics of AI. NBER Working Paper Series. 2024. DOI: 10.3386/w32487

https://lnkd.in/dSpumKxp This paper examines "The AI Layoff Trap", a phenomenon where competitive firms automate tasks faster than the economy can reabsorb displaced workers. While firms individualistically benefit from AI cost savings, they collectively erode the consumer purchasing power that sustains their own revenues. This creates a demand externality where a rational automation "arms race" leads to a deadweight loss that harms both workers and owners. The authors demonstrate that traditional solutions like upskilling, universal basic income, or worker equity fail to fix this market distortion. Instead, the research argues that only a Pigouvian automation tax can align private incentives with social welfare. Ultimately, the study suggests that increased competition and more capable AI actually intensify this destructive cycle.

Why does the same technology raise wages in one job, but lower them in another? A new theory of expertise, unveiled by David Autor and Neil Thompson. Last week, I wrote about David's Paris School of Economics/CEPR - Centre for Economic Policy Research keynote on expertise, but now the full working paper is out in National Bureau of Economic Research. Worth a read for not only labor economists, but any economist and scholar. Not all automation is created equal -- and neither are its effects on different occupations. David and Neil relate the effects with the role of expertise, referring to scarce and valuable human capital. A new framework, called the expertise model, helps explain why workers in jobs that look similarly “automatable” may experience radically different outcomes. Take two examples: accounting clerks and inventory clerks. Both were heavily exposed to automation in recent decades. Traditional economic models would predict similar effects -- declining wages or reduced labor share -- due to a loss of routine, codifiable tasks. But that’s not what happened. Instead: a. Accounting clerks’ wages have risen, while their employment declined. b. Inventory clerks’ wages declined, but employment rose. Why? Because automation eliminated inexpert tasks (like data entry) in accounting, pushing the role toward higher-skill, decision-oriented functions. Fewer people can do that work -- hence higher wages, but fewer jobs. In contrast, automation eliminated expert tasks in inventory roles (like pricing and flagging stock anomalies), reducing the skill barrier and opening the job to a broader labor pool -- hence lower wages, but more employment. This “expertise framework” is a powerful tool for understanding how technological change reshapes labor markets -- not just by eliminating tasks, but by changing who is qualified to do what remains. When asked about technological change, I used to think about things in terms of the task-based model, but now it's clear that we also need to consider how technology affects the optimal composition of expertise required within each job. And what's more, then how does AI affect the degree of augmentation versus automation of expertise within jobs? Organizations will need to respond accordingly (e.g., compensation, recruitment). #FutureOfWork #Automation #LaborEconomics #AI #WorkforceDevelopment #OccupationalShifts #HumanCapital #Productivity

Layoffs improve margins. But at scale, they can weaken the economy. At the firm level, cost-cutting is rational. Fewer employees → lower expenses → better short-term profitability. At the system level, it’s more complex. Jobs are not just a cost. They are demand. Wages paid by one company become: • Spending in another • Revenue elsewhere • Income that sustains the economy If layoffs become widespread, this cycle weakens. Today we’re seeing a pattern: • Companies optimizing costs • Slower hiring or workforce reduction • Increasing capital allocation toward AI and automation The issue is not technology. It’s the imbalance between productivity and income distribution. Historically: • Investment in factories and infrastructure created large-scale employment • Employment supported consumption and growth Now: • Investment in AI scales output • But generates fewer jobs relative to capital deployed This creates a structural risk: Rising productivity without proportional income growth. One possible policy idea (open for debate): Link employment to company scale. • Based on revenue or market cap, firms maintain a minimum employment threshold • If not, they face higher taxation or reduced incentives The intent is not to restrict efficiency, but to preserve demand in the system. Because if: • Fewer people earn → less spending • Less spending → weaker sales • Weaker sales → more cost cutting …it becomes a feedback loop. There’s also a fiscal dimension. If corporate tax bases shrink while unemployment rises: • Governments collect less from companies • Pressure shifts to individuals • Debt burdens increase over time Long-term, this can affect: • Consumption • Growth • Currency stability The core issue is simple: We need productivity growth. But we also need income distribution to sustain demand. Without both, efficiency gains may come at the cost of economic stability. Sharing this as a perspective, not a conclusion. How do we balance: • Innovation • Efficiency • And employment in the AI era?

AI is already impacting 93% of U.S. jobs, and its effects are outpacing expectations by a factor of 4.5x. The broader question for leadership today is no longer if, but how, we channel this transformative power to benefit all levels of society and the economy. Two perspectives stand out. In our TIME article, "AI Should Belong to Workers," my coauthors and I explored how AI disrupts traditional hierarchies by democratizing intelligence. Unlike previous technology waves, AI doesn't demand specialized technical expertise for adoption. Frontline employees — from HVAC technicians to nurses — are using AI-driven diagnostics to expand their leverage and decision-making. When workers shape AI applications tailored to their tasks, value creation accelerates closer to the work itself, enabling better wage leverage and upward mobility. (read here: https://lnkd.in/eSwqbDpT) Our "New Work, New World" research, spanning nearly 1,000 professions, found AI exposure has accelerated well beyond predictions — average exposure scores jumped 30% in just three years, when models forecasted a decade. The yearly rate of jobs impacted by AI has skyrocketed from 2% to 9% annually, underscoring why leaders must design pathways that empower workers to harness this opportunity at every level. (read here: https://lnkd.in/ekWPUxQE) In the Newsweek article, "When Capital Can Think, Who Pays?", we examined AI's fiscal misalignment: digital agents that augment or replace workers contribute nothing toward payroll taxes sustaining Social Security, Medicare, and unemployment insurance, while human labor bears these costs. Temporarily rebalancing this burden — lowering it on labor, raising it on automation — creates an AI-driven economy that rewards augmentation over displacement, much like prior revolutions that generated the 60% of jobs that didn't exist 80 years ago. (read here: https://lnkd.in/eUvMvZMK) AI's potential to create value for the U.S. economy already exceeds $4.5 trillion. But unless enterprise adoption and wider societal architecture move in tandem, these gains could concentrate in narrow economic bands. Leaders today have a choice: manage AI passively, reinforcing the inequities new technologies could correct — or reimagine how intelligence, tasks, and rewards flow through our organizations. The most important innovation of the coming decade may not come from AI itself. It will come from the deliberate systems we create to amplify every worker's potential and ensure technological progress fuels enduring human value.

"When automation removes the simpler tasks (as accounting software did for bookkeeping clerks), the remaining work becomes more specialized, wages rise, and fewer workers qualify. When it removes the harder tasks (as inventory management systems did for warehouse workers), the job becomes more accessible, employment expands, and wages fall. Same technology, opposite labor market outcomes, depending on which part of the job gets automated." - Alex Imas, The University of Chicago Booth School of Business from What Will Be Scarce? https://lnkd.in/eqdKHk2u

𝗧𝗵𝗲 𝗙𝗼𝗿𝗴𝗼𝘁𝘁𝗲𝗻 𝗥𝗼𝗹𝗲 𝗼𝗳 𝗘𝗻𝘁𝗿𝘆-𝗟𝗲𝘃𝗲𝗹 𝗝𝗼𝗯𝘀 𝗶𝗻 𝗮𝗻 𝗔𝗜 𝗘𝗰𝗼𝗻𝗼𝗺𝘆 AI promises massive productivity gains. But it may also be quietly eroding how expertise is built. As AI enables senior employees to do more on their own, many entry-level roles—the primary source of learning by doing—are disappearing. This matters because the most valuable workplace skills are often 𝘁𝗮𝗰𝗶𝘁: absorbed through experience, not taught in classrooms or manuals. According to a recent study, today’s rush to automate early-career work may be socially excessive. While automation boosts short-term productivity, it also disrupts the intergenerational transfer of tacit knowledge. The result is a trade-off: higher output now, but weaker skills in the next generation—ultimately slowing long-term economic growth and, in some cases, reducing overall welfare. The implications are not trivial. Even modest levels of AI-driven automation at the entry level could lower long-run U.S. per-capita growth by an estimated 𝟬.𝟬𝟱 𝘁𝗼 𝟬.𝟯𝟱 𝗽𝗲𝗿𝗰𝗲𝗻𝘁𝗮𝗴𝗲 𝗽𝗼𝗶𝗻𝘁𝘀 𝗮𝗻𝗻𝘂𝗮𝗹𝗹𝘆. Over time, that compounds into a meaningful economic drag. AI co-pilots offer a partial remedy. They can help workers who missed early learning opportunities catch up later in their careers. But they also introduce a new tension: if AI makes skill gaps easier to mask, it may reduce incentives for juniors to develop those skills in the first place. 𝗧𝗵𝗲 𝗺𝗲𝘀𝘀𝗮𝗴𝗲 𝗶𝘀 𝗹𝗼𝘂𝗱 𝗮𝗻𝗱 𝗰𝗹𝗲𝗮𝗿: 𝗔𝗜 𝗰𝗵𝗮𝗻𝗴𝗲𝘀 𝗻𝗼𝘁 𝗷𝘂𝘀𝘁 𝗵𝗼𝘄 𝘄𝗼𝗿𝗸 𝗶𝘀 𝗱𝗼𝗻𝗲, 𝗯𝘂𝘁 𝗵𝗼𝘄 𝗲𝘅𝗽𝗲𝗿𝘁𝗶𝘀𝗲 𝗶𝘀 𝗳𝗼𝗿𝗺𝗲𝗱. 𝗔𝗻𝗱 𝗴𝗿𝗼𝘄𝘁𝗵 𝗱𝗲𝗽𝗲𝗻𝗱𝘀 𝗼𝗻 𝗯𝗼𝘁𝗵. To capture AI’s full potential, policy, firms, and universities must protect and expand early-career learning—through mentorships, apprenticeships, practical education, and AI systems that complement junior roles rather than erase them. 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝘃𝗶𝘁𝘆 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝘀𝗸𝗶𝗹𝗹 𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝗻𝗼𝘁 𝗽𝗿𝗼𝗴𝗿𝗲𝘀𝘀—𝗶𝘁’𝘀 𝗯𝗼𝗿𝗿𝗼𝘄𝗲𝗱 𝗴𝗿𝗼𝘄𝘁𝗵. 𝗥𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲: Ide, Enrique. (2025). Automation, AI, and the Intergenerational Transmission of Knowledge. 10.48550/arXiv.2507.16078.