At a recent IAWP webinar, workforce professionals had the opportunity to hear from Ben Armstrong, Executive Director of the Industrial Performance Center at the Massachusetts Institute of Technology and co-leader of MIT’s Work of the Future Initiative. Armstrong leads research examining how technology affects productivity, skills, and workforce demand across industries including manufacturing, healthcare, finance, insurance, and logistics.
Drawing on research conducted with more than 50 companies studying the use of generative AI, Armstrong shared what organizations are actually experiencing as these tools enter the workplace. His message offered a useful counterbalance to the dramatic predictions often seen in public discussions about artificial intelligence and jobs.
Moving Beyond the AI Hype
Armstrong explained that many claims about AI assume sweeping labor market disruption. In practice, the changes are usually more incremental and depend heavily on how organizations adopt and integrate the technology.
“Technologies typically affect tasks within jobs, not entire occupations,” Armstrong noted.
This distinction is important for workforce professionals. Most jobs consist of many different responsibilities, and while AI may assist with certain tasks, other parts of the role continue to rely on human judgment, communication, and contextual understanding.
Lessons from Earlier Waves of Technology
Much of MIT’s research examines how past technologies—such as robotics, industrial automation, and information technology—have changed work. One consistent finding is that organizations often struggle to fully automate complex work processes.
Companies may successfully automate certain routine activities, but eliminating entire jobs is much more difficult. In many cases, workers remain essential because they provide flexibility, problem-solving ability, and expertise that automated systems cannot replicate.
Armstrong illustrated this idea with a simple example: the bread maker. Engineers were eventually able to automate parts of bread-making only after studying the tacit knowledge of experienced bakers. Even then, bread machines did not eliminate the profession. Technical capability alone does not determine whether work becomes automated.
How AI Is Changing Work Inside Organizations
MIT’s research identified several patterns emerging across industries. In some settings, workers are shifting from performing tasks manually to supervising automated systems and interpreting data. This trend is visible in sectors such as manufacturing, utilities, healthcare, and financial services.
At the same time, the need for workers who can diagnose and repair problems is increasing. As automated systems become more common, organizations rely heavily on people who can troubleshoot issues, interpret system failures, and restore operations.
Even in highly skilled professions often cited as vulnerable to AI, the reality is more complex. Radiology, for example, is frequently mentioned as a field likely to be replaced by algorithms. Yet the number of radiologists has continued to grow. Their work includes far more than pattern recognition; it also involves patient communication, medical judgment, and coordination within the healthcare system.
Adoption Is High, but Value Is Still Unclear
Another key finding from the research is the difference between AI adoption and AI value.
Many organizations have adopted generative AI tools quickly, often because employees were already experimenting with them. But high usage does not automatically translate into productivity gains. Companies are now asking a more difficult question: where do these tools actually create measurable improvements in performance or outcomes?
Early experimentation across industries shows that many pilot projects never scale successfully. Integrating new tools into existing workflows, building trust among employees, and measuring real results remain significant challenges.
The Risk of Skill Atrophy
Armstrong also raised concerns about how AI might affect learning. Emerging research suggests that while generative AI can help people complete tasks faster, it may also reduce how much knowledge they retain from the work itself.
If workers rely heavily on AI-generated outputs, they may gain efficiency but lose opportunities to develop deeper understanding. For training systems and educational institutions, this raises an important question: how to use AI as a support tool without weakening the expertise workers need.
Implications for Workforce Development
For workforce development professionals, Armstrong offered several practical insights.
First, it is important to focus on tasks rather than entire occupations when thinking about how AI will affect work. Many roles will evolve rather than disappear.
Second, domain expertise remains critical. Workers still need strong occupational knowledge in order to evaluate AI outputs, interpret results, and apply them effectively.
Finally, workforce professionals themselves should gain familiarity with these tools. Even when agencies place limits on their use due to privacy or security concerns, understanding how AI behaves—and where it succeeds or fails—can help professionals better advise job seekers and employers.
A Measured View of the Future
The broader message from MIT’s research is that artificial intelligence is neither an immediate labor market catastrophe nor a guaranteed productivity revolution. Like previous technological advances, its impact will depend largely on how organizations choose to implement it.
For workforce professionals, that means the challenge is not simply reacting to headlines about AI. It is helping workers, employers, and training systems make thoughtful decisions about how these tools are used—so that technology strengthens jobs rather than weakens them.