The White House is pushing ahead with its plans for government integration of artificial intelligence. Just recently, the Trump administration outlined directives for federal agencies, aiming to “pro-promotion” AI to influence how government functions in the future.
The central question remains: Will AI displace workers? More specifically, will many in the federal workforce see their roles change significantly after the latest developments in technology, or will it merely enhance productivity and open more opportunities?
Whether or not AI plays a role, technological advancements are reshaping employment landscapes constantly. Sometimes they eliminate positions, but they also create new ones. For instance, innovations in robotics and software have diminished the need for certain manufacturing and administrative jobs, while also generating new positions that demand higher skills.
A recent study published last month in a top economics journal sheds light on a familiar subject: AI’s impact on jobs.
This research focused on customer service agents at a Fortune 500 software company, where GPT-3-powered AI tools were deployed. This scenario, marked by high demands and frustrated customers, proved to be a fitting ground for such advanced AI support.
The results were striking. Overall productivity surged by 15%, with less experienced agents experiencing a 36% improvement. Essentially, those who needed help the most benefited significantly.
By addressing knowledge gaps in real-time, AI helped reduce the time spent on customer interactions and enabled staff to handle more inquiries per hour, while top performers saw little change in their outputs.
This data challenges the prevailing view that low-skilled workers are the first to suffer from new tech.
AI has seemingly leveled the learning curve, allowing for quicker skill acquisition. What typically took years can now be condensed into months, which, if this trend continues, may lead to gains for both businesses—through enhanced productivity—and workers—by acquiring valuable skills and better income potential.
The study also suggests that AI could help balance labor market disparities created by the first waves of automation.
For decades, technology advancements have trickled in slowly, particularly impacting mid-level jobs in areas like office management and manufacturing, which often left many stuck in low-wage roles without higher education. Meanwhile, professionals in creative or judgment-based roles appear more insulated from these shifts.
AI seems to hold the potential to widen access to real-time expertise, potentially expanding the number of workers eligible for decision-making roles. Picture a scenario where, instead of pushing for economic nationalism, AI helps rebalance the labor market, making skilled jobs a pathway to economic mobility again.
However, AI doesn’t necessarily devalue top experts. It might merely allow a broader base of individuals to engage in tasks that previously required extensive training. You might think of it as a way to democratize expertise—with AI serving as a practical mentor and research tool.
For instance, while AI might exceed the average radiologist in reading mammograms, it still struggles with various tasks, like comforting patients. Research indicates that while AI can perform functions across nearly all 950 occupations, it’s not flawless across the board.
According to David Autor, an economist at MIT, technology has historically created new job roles. Around 60% of today’s jobs didn’t even exist in 1940, which underscores how technology can continually reinvent human labor.
Much like ATMs didn’t eliminate the need for tellers—they simply shifted their roles toward customer service—well-designed AI can liberate workers from mundane tasks, allowing them to focus on critical thinking and creativity.
To ensure AI supports a robust labor market, policymakers must address the current tax biases against human capital.
Right now, businesses can only deduct training costs if they’re job-specific. But if they were allowed to deduct the full costs of training, similar to how they do with buildings or tech, it could encourage companies to reskill their workers for higher-value positions. These reforms could incentivize active reskilling, helping people advance their careers in an AI-driven economy.
Equally crucial is enhancing financial flexibility for individuals to invest in their own skill development. Universal Savings Accounts could offer simpler methods for low- and middle-income workers to save and utilize funds for training without the complexities of existing benefit systems.
That said, it’s important to remain aware of the pitfalls associated with AI. The key takeaway might be, how we shape our workforce through AI will greatly depend on our actions.
Directing AI to enhance human capabilities could foster wider employment opportunities and help reclaim the skilled workforce that has been impacted by past technological changes.
Liya Palagashvili is a senior researcher and director of labor policy at George Mason University’s Mercatus Center. Revana Sharfuddin is a predoc researcher at Mercatus and the author of a recent post on Substack.Trusting our humanity in the AI era.
