Key Takeaways

  • Cardinal Robert W. McElroy warned that AI and robotics could produce major workforce dislocation over the next two decades.
  • Labor research suggests AI will change far more jobs than it fully automates, placing greater pressure on training and job redesign.
  • Employers can incorporate worker participation, impact assessments and governance standards into AI deployment decisions.

Cardinal Robert W. McElroy used a Labor Day Mass in Washington to challenge business leaders, policymakers and labor representatives to address a difficult uncertainty surrounding artificial intelligence: whether technological progress will leave enough meaningful work for people.

“Will there be meaningful work left?” Cardinal McElroy asked during his Sept. 6 homily at the Cathedral of St. Matthew the Apostle. “And the answer to that is: No one knows.”

The Mass, organized by the Catholic Labor Network, brought together parishioners, workers and organized labor representatives. Cardinal McElroy framed employment as more than an economic transaction. Work, he said, allows people to create, contribute to communities and express part of themselves.

That perspective has practical consequences for enterprise AI strategies. If executives measure automation primarily through productivity, headcount and cost, they may overlook changes to career development, employee autonomy and institutional knowledge. An AI system can improve output while still weakening the pathway through which junior employees learn how a business operates.

The concern is becoming more immediate. Cardinal McElroy said recent discussions between U.S. cardinals and AI leaders in Silicon Valley included the combined workforce effects of AI and increasingly capable robotics. He predicted that the labor market would change radically over the next 20 years, with millions of workers no longer holding their current jobs within 10 years.

Research supports the expectation of broad disruption, although not necessarily mass job elimination. The International Labour Organization estimated in 2026 that about 24% of workers worldwide are potentially exposed to AI, while concluding that most jobs are more likely to be transformed than eliminated. In ASEAN, 22.9% of total employment, nearly 80 million workers, is in occupations with more than minimal generative AI exposure. Only 3.3% falls into the highest-exposure category.

Exposure is not the same as replacement. It can mean automated administrative work, AI-assisted decisions, revised performance expectations or a redistribution of tasks between experienced staff and entry-level employees. The OECD has also described AI’s labor-market impact as extending into wages, job design, skill requirements and the work environment. The European Training Foundation estimates that 40% to 60% of jobs could undergo significant task changes, while only a small proportion may be fully automated.

The August employment report offered a snapshot of an already uneven market. The U.S. economy added 162,000 jobs, and unemployment remained at a significant share. Manufacturing gained 16,000 positions and construction added 22,000. The information sector lost 23,000 jobs, including 8,000 in computing infrastructure, data processing, web hosting and related services. The report did not attribute those losses to AI, an important distinction when companies and policymakers assess early labor signals.

At the Mass, a representative of the International Brotherhood of Electrical Workers Local 26 pointed to skilled trades as careers that can lead into safety, business ownership and advocacy. A retired economist highlighted manufacturer partnerships with community colleges as one way to prepare people for jobs employers struggle to fill. Those models suggest reskilling works better when connected to identifiable positions, paid experience and regional demand rather than broad promises about future opportunities.

For employers, governance can bring workforce considerations into technology decisions earlier. The NIST AI Risk Management Framework and ISO/IEC 42001:2023 provide complementary approaches for identifying risks, documenting oversight and managing AI systems. Companies can extend those practices by assessing which tasks will change, who bears the risk of errors, how employees can challenge automated decisions and what training accompanies deployment.

Cardinal McElroy called for workers to have a place at the table as governments and businesses make those choices. That said, consultation alone may have limited value if it begins after procurement and restructuring decisions are settled. Earlier participation can help companies identify operational knowledge that models miss, preserve viable career ladders and share productivity gains more broadly. The central business question is not only which work AI can perform. It is what kind of work organizations still want people to have.