While many public discourses focus on concerns about advanced technology alternatives to human workers, new research from MIT Sloan Management School presents a different perspective. It simply identifies risky jobs from AI and emphasizes areas in which human expertise complements technological advances.
Read again: How can AI help businesses operate service and contact centres at a low cost?
The paper, “Epochs of AI: Human Machine Complementation in the Workplace,” provides a framework and set of metrics for human-intensive capabilities to assess tasks across all occupations and to better understand AI’s impact on the labor market. The authors are Roberto Rigobon, professor of business administration at MIT Sloan, and postdoc associate Isabella Loaiza.
“There tends to be a general story of robots coming for work,” Rigobon said. “I think it’s important to ask a variety of questions. It’s more important to look at human capabilities than AI capabilities and shift towards what technology can give us, rather than what technology can take away.”
Researchers have studied the statistical limitations of AI tools. AI is based on universal approximation functions, and these tools are known to perform poorly when data is biased or small, when extrapolation far from training data is required, and when moral dilemmas appear. From these flaws, the author focused on how humans dealt with these problems, creating the foundation of skills that complemented AI.
In this paper, we use three important metrics to assess tasks across different professions. Epoch index (including human abilities in five groups), applied risk score, and potential score. The acronym epoch is expressed as follows:
The existence of empathy and emotional intelligence, networking, connection, opinions, judgment, ethics creativity and imagination, imagination, vision, leadership.
Each of these categories includes the unique human ability to allow humans to work in areas where machines are restricted.
Metrics are used to assess whether humans are intensive tasks and whether they can automate or augment their occupations through technology. Automation involves direct transfer of tasks from humans to machines, but using machines in tasks increases worker productivity in that task or other tasks, and augmentation occurs when overall labor productivity increases. Therefore, augmentation must take into account interactions between tasks, whether pairs, clusters, or networks. Rather than simply acting as “partial automation,” enhancement allows humans to do things they couldn’t do before. For example, the introduction of advanced microscope tools has increased the ability of humans to work in the micro and nano world.
“Many research done in this field tends to use scores from there to extrapolate and see more generally detailed work activities,” Loaiza said. “We have focused specifically on the structure of tasks within the job or occupation to measure augmentation,” he said.
The findings suggest an increase in the amount of many important human-intensive tasks (tasks that cannot be performed entirely by machines) and the amount of human-intensive tasks, as well as an increase in the frequency with which workers performed these tasks between 2016 and 2024. It has a higher level of epoch functionality than previously existing tasks (pre-2024) and those that disappeared in 2024.
Latest Readings: We employ generation AI from proof of concept to production
Examples of tasks with high epoch levels include direct recruitment, placement, training and evaluation of architectural or engineering project staff. Such tasks include determining scientific or technical goals within the broad overview provided by top management and developing a detailed plan to achieve these goals. Examples of jobs that often include high levels of epoch skills such as creativity and empathy include emergency management directors, clinical psychologists, counseling psychologists, childcare workers, public relations experts, and film directors.
The study points to the need to invest in the development of workers’ epoch capabilities and benefit from helping workers become complementary and unsubstituted by AI and new technologies.
“We don’t intentionally call these (human skills) “soft” skills,” Rigobon said. “Hard” skills like solving mathematical problems are relatively easy to teach. It is much more difficult to teach people these important human skills and abilities, such as hope, empathy, creativity. ”
(To share your insights with us as part of your edited or sponsored content, please write at psen@itechseries.com)
