In the White House’s recent executive order, dubbed the “American AI Initiative,” there was a notable focus on Science, Technology, Engineering, and Mathematics (STEM) education and its role in securing American technological leadership. This followed from the Trump Administration’s earlier plan to increase STEM literacy. While it is certainly true that technical skills are lacking, these pivots may be harmful to sustainable growth.

It is undoubtedly the case that greater STEM competence increases the likelihood of scientific breakthroughs, new technologies, and medical advances, but none of these in isolation have an outstanding social impact. They need complements, from business, policy, and media, who are able to help spread and magnify their potential. A narrow focus on STEM alone does not enable these to arise. The fact that we have had so many new technologies appear in recent years, but they have only marginally contributed to increased productivity, indicates that there is a need for these complements.

In addition, policy that directs resources at accelerating the development of science and technology may simultaneously make society less adaptable to its consequences. Those trained in STEM, bar exceptional performers, usually do not have skills that are valuable across their entire working lives. Economists David J. Deming and Kadeem Noray at Harvard found that after a decade in the workforce, technological change and constantly improving scientific knowledge contributes to a significant decline in the earnings of those with engineering and computer science backgrounds. Younger and newly trained workers are able to do the work better and for less pay than experienced workers using outdated software or focusing on problems with declining societal interest. As technological innovation automates existing tasks and introduces new ones, investments that lock people into particular vocations make them less adaptable.

They interpret these findings as showing that the lack of STEM workers in the economy is less about increasing trained graduates but encouraging contextual skills. Raising STEM enrolment at universities then will likely be less effective in the long run than finding alternative means to increase the number of skilled workers. By focusing on learning skills, either on the job or through online education, rather than commitments to formal education, the risk of credential inflation pushing down wages is reduced. Economists Seth Benzell and Erik Brynjolfsson at MIT argue in a recent paper that scarce complements are crucial drivers of firm success in a digitized economy. Raising the skill level of the average worker, they posit, would provide them with more of what is already abundant, increasing competition for middle-income jobs, driving down wages.

Increasing technical skills narrowly then would not benefit workers or lead to long-run improvements in innovation, but this does not mean it is not an important task. Rather than raising the average level of technical skills present in an economy, the focus should be on increasing the diversity of talent. Combining technical training with communication skills, policy expertise, management, ethics, or a whole host of other difficult to automate, yet essential contributors to innovation, would provide a far more sustainable growth strategy.

This versatile skill base is increasingly demanded by the quaternary sector, which increasingly dominates post-industrial economies, and places radical demands on education systems, and it is abundantly clear that prioritizing certain fields over others is not enough to resolve these issues. An interdisciplinary approach to innovation can be seen in Canada, where organizations like the Vector Institute support not only core AI research, but AI-related fields, and work with industry to raise technical literacy among managers.

Critics of the Trump administration’s push on STEM have noted that many of the proposals put forth have been rolled out across the country over the last decade, as schools have experimented with more personalized learning tools and education technologies. Organic responses that engage relevant stakeholders are far more adaptable than strategies locked in from the top-down.

While it is still unclear what the American AI Initiative will entail in practice or whether the White House’s STEM strategy will be carried out, it should nonetheless be apparent that a holistic focus is essential to delivering sustainable economic growth.