Can AI help prevent loss of future scientists?

By Mehwish Khokhar

(The writer is an American Chemical Society professional, originally from Lahore)

The Hidden Cost of STEM Attrition

Every year, millions of students begin their journey into science, technology, engineering, and mathematics (STEM) education. Yet many never complete that journey. According to data from the National Science Foundation, only about 55% of students who begin STEM pathways ultimately complete a STEM credential or degree, creating a significant challenge for educational institutions, employers, and policymakers seeking to strengthen the future scientific workforce.

The need for solutions has become increasingly urgent. The U.S. STEM workforce now exceeds 36 million people and represents approximately one-quarter of the nation’s workforce. At the same time, STEM occupations are projected to grow faster than non-STEM careers, driven by expanding industries such as artificial intelligence, biotechnology, clean energy, advanced manufacturing, and semiconductor production. As the United States continues investing in technological innovation and scientific competitiveness, retaining STEM talent has become as important as recruiting it.

Against this backdrop, Khokhar’s recently filed patent application introduces a predictive framework designed to help educational institutions identify students who may be at risk of leaving STEM pathways before disengagement occurs. By combining educational analytics, workforce intelligence, and personalized intervention planning, the proposed system aims to provide institutions with actionable insights that support student persistence, workforce readiness, and long-term career success.

The patent application outlines a framework intended to:

• Identify early indicators associated with future STEM attrition before traditional academic measures reveal a

• Recommend personalized interventions such as mentorship opportunities, research experiences, internships, scholarships, and career-development programs.

• Forecast which interventions may be most effective in improving student persistence and workforce readiness

• Provide universities, workforce organizations, and policymakers with data-driven insights to strengthen STEM talent pipelines and address future workforce

Rather than relying solely on historical outcomes, the invention seeks to help institutions move toward a more proactive and evidence-based approach to student retention and workforce development.

From Workforce Development to Innovation

The idea emerged from Khokhar’s professional experience at ACS, the world’s largest scientific society. Through her work, she has supported initiatives connecting students, graduate researchers, postdoctoral scholars, educators, and industry professionals through mentorship, professional development, and workforce engagement opportunities. Her efforts have contributed to programs reaching more than 15,000 students across over 60 universities, providing her with direct insight into the challenges facing future scientists and engineers.

Over time, Khokhar observed recurring patterns among students who struggled to remain engaged in STEM fields. While institutions collected significant amounts of information regarding academic performance, participation, and student outcomes, much of that information remained fragmented and underutilized. These observations ultimately led to a larger question: could technology be used to predict future attrition risks before talented individuals permanently exit STEM pathways?

That question became the foundation of her patent application.

Reimagining the Future of Scientific Talent

The invention has also generated interest among educators who work closely with STEM students and workforce-development initiatives. Dr. Naeem Asad, former professor at Forman Christian College, Lahore, believes the concept addresses a challenge that extends far beyond a single institution.

“Universities have long struggled to identify which students may be at risk of leaving STEM pathways before meaningful intervention becomes possible,” he said. “What makes this concept particularly interesting is its attempt to combine predictive analytics, workforce intelligence, and student-development data into a single framework. If successfully implemented, it could provide educational institutions with a more proactive approach to supporting future scientists and engineers.”

Through her work in STEM workforce development, Khokhar has focused on strengthening pathways that connect students, researchers, and early-career professionals with opportunities in science and technology. Her newly filed patent application represents an effort to transform STEM retention from a reactive challenge into a predictive science, with potential applications for universities, workforce development organizations, scholarship programs, scientific societies, and government agencies seeking to strengthen the next generation of scientific talent.

As conversations around artificial intelligence, workforce readiness, and scientific competitiveness continue to evolve, innovations aimed at understanding and strengthening the STEM pipeline may play an increasingly important role in shaping the future of education and innovation.

From Lahore to one of the world’s largest and most prestigious scientific organizations, Mehwish Khokhar’s career reflects the increasingly global nature of scientific leadership. The Pakistan-born STEM workforce professional, now working at the American Chemical Society (ACS), has filed a United States patent application with the U.S. Patent and Trademark Office (USPTO) for a technology designed to help educational institutions identify and address STEM attrition before students leave scientific pathways.