Living Faithfully in an Economic Revolution, Part II: AI’s Impact on Productivity and Labor

In Part 1, we looked back at how the Industrial Revolution fundamentally enhanced our capacity to turn physical materials into consumer goods, bringing economic growth and societal growing pains.

The AI Revolution may similarly enhance our capacity to turn information into insights. How this will unfold remains uncertain, but economics provides useful frameworks for assessing the potential effects of AI on what we produce and the work we do. I’ll draw upon these frameworks to sketch out what the future could look like and close with how we might respond as Christians.

What is AI from an economic perspective?

“What a computer is to me is it’s the most remarkable tool that we’ve ever come up with, and it’s the equivalent of a bicycle for our minds.” – Steve Jobs

Humans aren’t especially strong or fast compared to the rest of the animal kingdom. Yet through increasingly complex tools, we’ve become unmatched in our ability to reshape the physical world. Steve Jobs once described the computer as “a bicycle for our minds.” If Jobs represents the era of bicycles for the mind, perhaps we are now entering the era of the Model T for the mind. ChatGPT has become ubiquitous almost overnight.

But AI is bigger than generative AI or large language models. More broadly, AI refers to computer systems that can classify, predict, and analyze by learning patterns from large amounts of data. Economically speaking, AI is best understood as a general purpose technology[1]—like the steam engine, electricity, or the computer—whose effects spill across nearly every industry rather than being confined to a single sector.

What makes AI potentially different from previous technologies is its domain. Past innovations helped us move around, move goods, and move information, often complementing and multiplying human intelligence. But AI raises the possibility of automating intelligence itself. If machines become capable of performing a wide range of cognitive tasks, and perhaps eventually physical tasks when paired with robotics, the scale of economic change could be enormous.

Economists often represent production using production functions, which describe how inputs like capital and labor generate output. Labor economists also use task-based models that describe occupations as bundles of tasks. Building on these frameworks, Daron Acemoglu models AI as a shift in the productivity of capital (machines) across tasks previously performed by labor.[2] The key question becomes: how many tasks can AI perform at lower cost than humans, and how much more productive will it be?

How might AI impact the economy?

“It’s tough to make predictions, especially about the future.” – Yogi Berra

The economic literature on AI is currently exploding. Researchers are building theoretical models to explore possible futures, and early empirical estimates suggest a wide range of outcomes. Much depends on how quickly AI improves, how widely it is adopted, and whether it mainly complements human workers or replaces them.

Productivity Effects

The first question is about economic growth: how much more productive will AI make the economy as a whole? A recent report from economists at the Federal Reserve Bank of Dallas illustrates the uncertainty by presenting several possible paths for real GDP per capita over time.

Historically, the U.S. economy has grown at around 2% per year since 1870, a remarkably steady trend even through the rise of computers and the internet. Will AI simply sustain this trend—or, like the Industrial Revolution did in the late 1800s, will it break it?

Scenario 1: A “Benign Singularity”
In the most dramatic scenario, artificial general intelligence (AGI)—AI that can perform essentially any human cognitive task[3]—could usher in a period of much faster economic growth.[4] If AI becomes a close substitute for human labor across a wide range of tasks, growth could accelerate sharply.[5] In the extreme, if AI automates research and development and contributes to improving itself, technological progress could compound at an unprecedented pace.[6] This scenario is probably unlikely (at least any time soon)._But even a small probability would have enormous implications.[7]

Scenario 2: A Large Increase in Growth
More moderate forecasts predict a meaningful but not world-transforming productivity boom. For example, Goldman Sachs has estimated that AI could raise GDP by roughly 15% over a decade, and McKinsey predicts similar gains.[8] If these projections are right, AI could raise growth rates by one to several percentage points per year—enough to noticeably change living standards.

Scenario 3: A Modest Productivity Boost
Other economists are more cautious. Acemoglu, for example, uses estimates of how many labor tasks are realistically exposed to automation and how much AI reduces costs in those tasks.[9] Under these assumptions, he projects a relatively modest productivity gain—on the order of less than one percentage point over the next decade. His approach is conservative, largely reflecting what AI can reliably do today, and it does not account for AI-driven innovation or large-scale organizational changes within firms.

Even revolutionary technologies may take a long time to reshape the economy. Many production processes require complementary innovations and institutional adjustments before productivity gains show up in the aggregate data.[10] And adoption happens when AI is profitable and reliable, not merely when it is possible.

In that sense, there is a scenario where AI could face a fate similar to the computer. Robert Solow famously remarked: “You can see the computer age everywhere but in the productivity statistics.”

Labor Market Effects

Productivity gains sound great—but what happens if many human skills become obsolete? What if AI makes the economy far more productive, but the rewards disproportionately flow to the owners of capital rather than to workers? AI’s labor market effects are closely tied to its productivity effects, but even under the same growth scenario, very different outcomes are possible.

In Acemoglu’s framework, AI can raise output through four channels:

1. automation (machines perform tasks previously done by humans),
2. task complementarity (AI makes humans more productive in certain tasks),
3. deepening of automation (AI improves tasks that were already automated), and
4. new tasks (AI creates new forms of work or enables new industries).

All four channels can raise economic output, but automation is the one most likely to reduce labor demand and put downward pressure on wages. Korinek and Suh show in a theoretical model that wages could continue rising, or fall sharply, depending on how quickly automation expands and how many tasks remain difficult to automate.[11]

It is tempting to assume that AI will always complement rather than replace human labor. But AI systems continue to improve rapidly, and tools that once generated text are increasingly able to execute complex workflows, interact with software, and complete multi-step tasks.[12] If the set of automatable tasks expands far enough, AI could eventually become a close substitute for human labor across many occupations.

At the same time, humans are not simply bundles of tasks. Humans are moral agents, embodied creatures, and relational beings with purposes that cannot be reduced to productivity. People will still value human presence, human trust, and human creativity[13]—especially in areas like caregiving, teaching, ministry, leadership, and art. Even in a highly automated economy, human judgment will remain central in deciding what goals to pursue, what problems to solve, and what values should govern the use of powerful tools.

For these reasons, AI is unlikely to eliminate human work entirely. But it may radically reshape which skills are valuable, how labor is compensated, and how widely the gains from productivity are shared.

AI to our Detriment

It seems strange to me that humans have been imagining AI-caused dystopias in movies and books for so long, and yet now we’re finally on the precipice of an AI revolution. As long as humans are manning the controls in a reasonable way, I don’t expect robots to suddenly turn against us and kill us all. But powerful technologies can be used for good or evil, and AI will be no exception.

At a societal level, AI could dramatically expand the power of governments and corporations. What could a surveillance state do with facial recognition, drones, and predictive algorithms? And at a fundamental level, machine learning is powerful precisely because it can match patterns in complex, non-transparent ways. Automation becomes more tempting as we give the machine more control—but there is real danger in handing over the wheel too quickly.

Less dramatically, we’ve already seen how AI can be used to capture attention, distort truth, and manipulate behavior. Social media companies have become extremely effective at learning what keeps people scrolling. Unfortunately, our attention is not always drawn to what is true, beautiful, or good for us. One experiment found that while college students were willing to pay around $50 per month for TikTok or Instagram, they would also be willing to pay around $20 per month to have themselves and their entire social network leave the platform together.[14] Social media is profitable, but it may also be a trap we wish we could escape. If our own preferences reveal negative welfare effects, Christians should take seriously the possibility that an AI-driven attention economy may amplify temptations that we are already prone to embrace.

AI could be used to develop a new deadly virus and trigger a pandemic. Educators worry that AI could become a crutch that prevents students from learning how to think. Add in deepfakes, misinformation, cybersecurity risks, and pornography, and it is not hard to imagine how AI could be used for evil as well as good.

How can we respond?

Individually, I think Christians should stay curious, on guard, humble, and hopeful as we consider how to use AI in obedience to the cultural mandate and the great commission. We shouldn’t use AI in ways that replace our humanity (our embodiment, relationships, moral responsibility, and creativity) with something artificial. And we obviously shouldn’t participate in using AI to exploit others, nor let ourselves be drawn into sin by the traps others may set for us.

Societally, we need policies and institutions that mitigate the harms AI could bring: exploitation through surveillance and manipulation, widening inequality through automation, and a growing inability to establish trust in public life. The last decade has taught us that in a cacophonous online public square shaped by social media, we must speak truth and rebuild trust in our institutions.

The Industrial Revolution brought massive economic and social change. Although it eventually helped lift millions out of poverty, it also created urban working classes that suffered terrible conditions and meager wages. This crisis became known as “the social question,” and it demanded a response from the church. In the Catholic tradition, Catholic social thought sought to promote a just society by upholding human dignity and the common good through solidarity with the poor, protection of workers, and the principle of subsidiarity. In the neo-Calvinist tradition, Kuyper argued that biblical Christianity offered the foundations for a distinct alternative to both laissez-faire individualism and socialism. He called for government to protect the rights of labor, not just the rights of capital, and for a limited social safety net that provides aid for the poor when private charity fails to meet basic needs.[15]

Perhaps we will need a new Reformed social thought in response to the social questions raised by the AI Revolution. How do we uphold truth in an information economy fixated on entertainment, attention, and consumerism? How do we protect the rights of workers while allowing for the productivity growth that comes when AI takes over certain tasks? How do we make sure the economic returns to AI aren’t captured by the few at the expense of the many? What does sphere sovereignty look like when AI facilitates the centralization of power in a tech company that controls our information and shapes how we connect with other people?

These are difficult questions. But the answers will likely involve some regulation of tech companies and AI use, institutions that protect truth and rebuild trust, and local communities that choose to preserve forms of human-centered work rooted in relationship. A Reformed social thought for the AI era should apply Kuyper’s vision of sphere sovereignty and limited power to the challenges raised by the automation of intelligence, seeking to uphold human dignity, protect vocation, and restrain exploitation in an economy increasingly shaped by AI.

Ultimately, no one knows how AI will change our economy. But how we choose to use these powerful tools is in our hands, and we can rest secure knowing that we and our ever-changing world remain in God’s hands.

Footnotes:

[1] Bresnahan, T. F., & Trajtenberg, M. (1995). General purpose technologies ‘Engines of growth’?. Journal of econometrics, 65(1), 83-108.

[2] Acemoglu, D. (2025). The simple macroeconomics of AI. Economic Policy, 40(121), 13-58.

[3] Some experts believe we may already have reached AGI: Chen, E. K., Belkin, M., Bergen, L., & Danks, D. (2026). Does AI already have human-level intelligence? The evidence is clear. Nature, 650(8100), 36-40.

[4] Korinek, A., & Suh, D. (2024). Scenarios for the Transition to AGI (No. w32255). National Bureau of Economic Research. Davidson, T. (2021). Could advanced AI drive explosive economic growth. Open Philanthropy, 25.

[5] Trammell, P., & Korinek, A. (2023). Economic growth under transformative AI (No. w31815). National Bureau of Economic Research.

[6] Davidson, T., Halperin, B., Houlden, T., & Korinek, A. (2026). When Does Automating AI Research Produce Explosive Growth?.

[7] Davidson says, “My framework finds a ~15% probability of human-level AI by 2080.” Davidson, Tom, Could Advanced AI Drive Explosive Economic Growth?," Open Philanthropy, June 2021. https://www.openphilanthropy.o...

explosive-economic-growth/.

[8] https://www.mckinsey.com/capab...

[9] Acemoglu, D. (2025). The simple macroeconomics of AI. Economic Policy, 40(121), 13-58.

[10] Charles Jones and co-authors have shown how models with essential complementary tasks lead to a growth path that is governed by the “weak links” in the process: Jones, C. I., & Tonetti, C. (2025). Past Automation and Future AI: How Weak Links Tame the Growth Explosion. Working paper, Stanford GSB and NBER.

[11] Korinek, A., & Suh, D. (2024). Scenarios for the Transition to AGI (No. w32255).

[12] Claude Code is an important example of advances in generative AI. https://www.oneusefulthing.org...

[13] Raj, M., Berg, J. M., & Seamans, R. (2026). The artificial intelligence disclosure penalty: Humans persistently devalue AI-generated creative writing. Journal of Experimental Psychology: General.

[14] Bursztyn, L., Handel, B., Jiménez-Durán, R., & Roth, C. (2025). When product markets become collective traps: The case of social media. American Economic Review, 115(12), 4105-4136.

[15] Abraham Kuyper, “The Social Question and the Christian Religion,” in On Business & Economics, ed. Peter S. Heslam, Abraham Kuyper Collected Works in Public Theology (Bellingham, WA: Lexham Press, 2021), 11:169–229.