New research from the MIT Sloan School of Management challenges the prevailing belief that artificial intelligence is poised to replace large swaths of the workforce, suggesting instead that AI will more often complement human labor by enhancing productivity in ways machines alone cannot achieve.
The study, titled “The EPOCH of AI: Human-Machine Complementarities at Work,” proposes a new framework for understanding the relationship between human capabilities and machine learning tools. Rather than focusing on which jobs are most at risk, the authors emphasize the unique strengths humans bring to the workplace—especially in areas where AI struggles.
“There tends to be a prevailing narrative that robots are coming for jobs,” said Roberto Rigobon, co-author of the study and professor at MIT Sloan. “We think it’s important to ask different questions — looking more at human capabilities than AI capabilities.”
A new framework for evaluating human-intensive tasks
The research introduces the EPOCH index, which measures five categories of human capabilities: Empathy and Emotional Intelligence; Presence, Networking, and Connectedness; Opinion, Judgment, and Ethics; Creativity and Imagination; and Hope, Vision, and Leadership. These traits, the researchers argue, are essential in jobs where AI tools are least effective—such as in situations requiring ethical reasoning, extrapolation beyond existing data, or human-centered judgment.
The study also applies two additional metrics: a risk-of-substitution score and a potential-for-augmentation score. While substitution refers to tasks that can be fully handed over to machines, augmentation occurs when technology enhances a human worker’s productivity—allowing people to do more, or to do things they previously couldn’t.
“Augmentation allows us to consider how tasks interact within occupations,” said co-author Isabella Loaiza, a postdoctoral associate at MIT Sloan. “We focused specifically on tasks and the structure of those tasks, not just general job categories.”
Empirical trends show rise in human-intensive work
Using data from O*NET, a comprehensive U.S. labor database maintained by the Bureau of Labor Statistics, the researchers found a measurable increase in the number and frequency of human-intensive tasks performed between 2016 and 2024. Newly added tasks in 2024 exhibited higher EPOCH levels than those that were removed, pointing to a shifting labor market that increasingly values uniquely human skills.
Examples of roles with high EPOCH scores include emergency management directors, clinical psychologists, childcare providers, and public relations specialists—jobs that require leadership, empathy, ethical judgment, and creativity.
The findings underscore the importance of investing in human development alongside technological innovation. The researchers argue that supporting workers in cultivating EPOCH capabilities will help ensure they remain essential, rather than replaceable, in an evolving labor market.
“These are not ‘soft’ skills,” Rigobon said. “They are hard to teach and absolutely critical to the future of work.”
