Humans, Nature, and Machines: Will AI Create More Jobs Than It Replaces?

By Richard Sebaggala (PhD)

Economists have long debated whether new technologies create more jobs than they destroy. Each industrial revolution, from steam engines to electricity, sparked fears of mass unemployment, only for new industries and occupations to emerge. Artificial intelligence, however, feels different. It does not only automate physical tasks; it reaches into the cognitive space once thought uniquely human (Brynjolfsson & McAfee, 2014).

So far, the evidence suggests AI is not sweeping workers aside in large numbers. Instead, it is altering the composition of work by reshaping tasks rather than eliminating whole professions. Coders now refine AI-generated drafts instead of writing from scratch. Paralegals summarize less case law manually. Marketers polish content rather than produce the first draft. In this sense, AI resembles a new species entering an ecosystem. It does not destroy the entire environment at once but gradually reshapes niches and interactions (Acemoglu & Restrepo, 2019).

Where AI adds the most value is in partnership with people. In chess, teams of humans and AI working together often beat both the best human players and the best AI systems. The same pattern is emerging in business, law, and research: AI accelerates analysis and routine drafting, while humans provide judgment, context, and values (Big Think, 2025). I have seen this in my own work as a researcher. Recently when reviewing a colleague’s draft paper, I began by reading it closely and noting my own independent observations arising from my rich research experience. I realized the paper seemed to have too many objectives mentioned in abstract, introduction and in the conceptual framework, the moderating role was not reflected in the title but rather smuggled in the theoretical discussions and methodology, and the case study design did not align with the quantitative approach. These were my own reflections, grounded in my reading. Only afterwards did I turn to ChatGPT, asking it to check the validity of my comments, highlight further weaknesses, and frame the feedback in a structured way. The model confirmed my insights, sharpened the phrasing, and suggested revisions. In that process, the AI acted as a sparring partner rather than a substitute. My reasoning stayed intact, but my communication became clearer. This kind of human–machine cooperation illustrates why complementarities matter more than simple substitution.

I have also seen this dynamic in data analysis. When I begin with clear objectives and a dataset, AI tools can be very useful as a starting point. They can suggest methods for analysis, highlight possible weaknesses, and even recommend additional checks such as sensitivity tests or robustness tests. Some of these insights might have taken me much longer to discover on my own, and in some cases I might not have uncovered them at all. Yet the value lies not in letting the tool run the entire analysis, but in using its suggestions to sharpen my own approach. I have discovered that if you are proficient in data analysis using Stata, as I am, you can allow AI tools such as ChatGPT, Avidnote, or Julius to run analysis in Python, while staying in control by asking the AI to generate Stata do-files for each analysis. Since I already have the data, I can validate the results in Stata. The efficiency gains are significant: less time spent on routine coding, more time to ask deeper questions, and occasional exposure to advanced methods that the AI suggests from its wider knowledge base.

 

Nature reinforces the point. Disruption is rarely the end of the story. When new species enter an ecosystem, some niches disappear, but others open. Grasslands gave way to forests. Forests gave way to cultivated fields and cities. The same is true of labor markets. AI is closing some roles but creating others such as prompt engineers, AI auditors, ethicists, and data curators. The central economic question is not whether niches vanish, but whether workers are supported in adapting to new ones. Without adaptation, extinction occurs not of species, but of livelihoods (Acemoglu & Restrepo, 2019).

Some commentators imagine a post-work society, where intelligent machines carry most productive effort and people focus on creativity, care, or leisure. Keynes (1930) once speculated that technological progress would eventually reduce the working week to a fraction of what it was. More recent writers describe this possibility as cognitarism, an economy led by cognitive machines. Yet history shows that transitions are rarely smooth. Without preparation, displacement can outpace creation. That is why policy choices matter. Retraining programs, investments in AI literacy, experiments with shorter workweeks, and social safety nets can soften shocks and broaden opportunity. Just as ecosystems survive through diversity and resilience, economies need deliberate institutions to spread the benefits of transformation.

AI, then, is powerful but not destiny. Like natural forces, it can be guided, shaped, and managed. The real risk lies not in the technology itself but in neglecting to align human institutions, social values, and machine capabilities. If we approach AI as gardeners who prune, plant, and tend, we can cultivate a labor ecosystem that grows new abundance rather than fear. If we fail, the outcome may be scarcity and division.

History suggests that technology does not eliminate work; it transforms it. The challenge today is to ensure that transformation is inclusive and sustainable. Human ingenuity, like nature, adapts under pressure. Machines are the newest force in that story. The question is not whether AI will take all jobs, but whether we will design the future of work or leave it to evolve without guidance. My own practice of drafting first and using ChatGPT second reflects the broader lesson: societies must take the lead, with AI as an assistant, not a replacement.

References

Acemoglu, D., & Restrepo, P. (2019). Automation and new tasks: How technology displaces and reinstates labor. Journal of Economic Perspectives, 33(2), 3–30. https://doi.org/10.1257/jep.33.2.3

Big Think. (2025, September). Will AI create more jobs than it replaces? Big Think. https://bigthink.com/business/will-ai-create-more-jobs-than-it-replaces/

Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. W. W. Norton & Company.

Keynes, J. M. (1930). Economic possibilities for our grandchildren. In Essays in persuasion (pp. 358–373). Macmillan.