The rapid advance of artificial intelligence has revived an old debate: will automation destroy more jobs than it creates or trigger a productivity boom large enough to lift wages and living standards?
I approach that question by examining three data points. First, where job losses are already emerging. Second, which new roles are gaining traction? Third, what actions governments and companies must take to tilt the balance toward inclusive growth.
Related: Geoffrey Hinton: These Jobs Will Be Replaced Due to AI
Where job losses are most likely
Generative AI can now draft text, parse legal documents and handle routine customer queries in seconds. A March 2023 report from Goldman Sachs Research estimates that up to 300 million full-time jobs worldwide could be automated as the technology matures, with two-thirds of current occupations exposed to some degree of task change. The study highlights three sectors with the highest risk.
Contact centers: An NBER working paper that tracked 5,000 agents at a U.S. software firm found that an AI copilot raised average productivity by 14% and helped novice workers catch up to veterans. Management interpreted the gain as an incentive to consolidate teams and suggested that entry-level service jobs are vulnerable.
What about legal support? Research from the University of Pennsylvania and OpenAI shows that legal clerks and paralegals rank in the top decile for exposure to large-language-model tools, which means that at least half of their core tasks can be automated. Early adopters such as Allen & Overy report that generative AI drafts first-pass memos in minutes, trimming billable hours for routine work.
One more word about content services. The World Economic Forum’s Future of Jobs Report 2023 projects a net loss of 16% in clerical and secretarial roles by 2027 as generative tools take over copy editing and basic design. Media companies experimenting with AI headlines and video summaries confirm that fewer junior editors are needed per project.
Related: Walmart CEO: AI Will Transform ‘Literally Every Job’
Where new jobs are appearing
The same WEF survey predicts that demand for AI and machine-learning specialists will rise by 40% through 2027, adding nearly one million positions. Data analysts, cybersecurity experts and prompt engineers also make the growth list.
IBM’s August 2023 Institute for Business Value study reinforces the shift, noting that 40% of the global workforce will need reskilling within three years because of AI adoption.
In other words, technology is not merely killing roles; it is redefining skill mixes inside most professions.
The inequality challenge
Job displacement rarely hurts all workers equally. Goldman Sachs finds that white-collar, higher-paid jobs carry the greatest exposure, yet low-income employees have fewer resources to retrain. Without policy intervention, the gap may widen.
McKinsey Global Institute warns that wage polarization will intensify unless reskilling keeps pace, even in optimistic growth scenarios. I believe that the political sustainability of AI adoption will hinge on visible pathways for at-risk workers to pivot into the new roles.
This technological shift forces us to confront a deeper, more uncomfortable question: what is the true purpose of an economy? Is its ultimate goal simply maximum efficiency, or is it to provide meaningful livelihoods for the people within it?
The danger isn’t just unemployment figures, but a hollowing out of the early and mid-career rungs that have traditionally allowed people to build skills, confidence and a sense of professional identity. We risk creating a world where a hyper-skilled elite designs and manages the AI, while a larger segment is left with precarious gig-work, constantly retraining for roles that may themselves become obsolete in a few years.
This isn’t just an economic problem; it’s a social stability problem in the making.
What governments should do
Regulation, reskilling and adaptation are not mutually exclusive. The European Union’s draft AI Act imposes transparency and risk-management duties on companies that deploy advanced models, aiming to build public trust without freezing innovation.
In my view, combining guardrails with broad-based training subsidies spreads both the risks and rewards of automation.
How companies can respond
Several multinationals have started large-scale reskilling. Microsoft’s Global Skills Initiative has trained over nine million people in digital competencies since 2020, with new modules on generative AI released this year. Amazon offers its employees free enrollment in its Machine Learning University, citing internal data that shows reskilled workers move into higher-paid technical roles within two years.
Companies also have strategic reasons to upskill rather than lay off. IBM’s CHRO tells Harvard Business Review that retraining is far more cost-effective than external hiring for advanced analytics roles. Internal mobility preserves institutional knowledge and reduces onboarding risk.
The cost of inaction
Ignoring workforce transition can backfire. A 2024 PwC survey of 345 investors and analysts showed that almost half of the respondents believed that growth required moderate to significant investment into AI and AI-powered tools. Productivity gains require human capital capable of steering and validating model outputs. I argue that layoffs without retraining may lift short-term earnings but erode long-term competitiveness. That is, the best way forward is retention, retraining and investment.
Related: The Leadership Style That’s Winning in the AI Era
A balanced path forward
My conclusion is pragmatic. AI will eliminate certain roles in call centers, legal support and content services. It will also create high-value jobs in data science, cybersecurity and human-centric design. Governments should set clear usage rules and fund lifelong learning, while companies should treat worker redeployment as a strategic asset, not a compliance chore. If both sectors act in tandem, we can steer the current wave of automation toward a new industrial revolution rather than a mass-unemployment crisis.