While working at NASA’s Jet Propulsion Laboratory (JPL) in the 1990s, I led the development of the “camera-on-a-chip” — a breakthrough that replaced bulky camera components with a tiny semiconductor image sensor. Today, the technology my team and I pioneered — known as the complementary metal-oxide-semiconductor (CMOS) image sensor — powers the cameras in smartphones, laptops, autonomous vehicles, and countless other products. Roughly seven billion CMOS image sensors are now produced each year.

But inventing the camera-on-a-chip was only the beginning. Turning a government-funded technical advance into a product used by billions required something just as important — a way to move that technology from the laboratory into the marketplace.

That’s what Congress sought to provide when it passed the Bayh-Dole Act in 1980. By allowing universities, research laboratories, nonprofit organizations, and small businesses to retain patent rights to inventions arising from federally funded research, the law created a framework for technology transfer that has helped turn countless discoveries into commercial products.

But today, that system faces new uncertainty. Several federal agencies are considering reforms to the rules governing research grants and intellectual property. If enacted, those proposals could make the path from early-stage research to commercialization harder to navigate successfully. Before policymakers make such changes, they should consider what the current approach has made possible.

My own experience shows why caution is warranted. Before Bayh-Dole took effect, the government generally owned taxpayer-funded inventions developed in university laboratories. Yet, by 1980, fewer than 5% of the roughly 28,000 patents the government owned had been licensed for commercial use. As a result, many promising inventions never became products that could benefit the public.

The Bayh-Dole Act changed that. And the results have been remarkable. Since Bayh-Dole became law, university technology transfer has helped launch more than 21,000 technology startup companies, developed over 200 new drugs and vaccines and contributed nearly $2 trillion to the U.S. economy.

After our team developed the camera-on-a-chip at JPL, Dr. Sabrina Kemeny and I founded Photobit Corporation to bring the technology to market. We received an exclusive license to our invention from the California Institute of Technology, which operates JPL and owned the patent rights under the Bayh-Dole Act.

That exclusive license gave us the assurance we needed to attract investment and move the technology beyond the research lab. From there, our company worked to refine the sensor and persuade electronics companies to adopt an unfamiliar alternative to the dominant imaging technology.

Photobit wasn’t the end of the story for me. As a professor at Dartmouth and later as co-founder of Gigajot Technology Inc., I’ve continued developing new imaging technologies and working to bring those innovations to market. I have helped dozens of faculty and PhD students commercialize their own inventions for the benefit of society. Successful startups benefit taxpayers and our government through employment, improved health outcomes, and public safety. Larger companies often acquire these startups once the underlying technology and market have been derisked, adding further value to our economy.

None of this can happen without Bayh-Dole.

That success demonstrates what’s possible when public research, entrepreneurial initiative, and strong intellectual property protections work together. Bayh-Dole helped give researchers, universities, startups, investors, and established companies a shared interest in transforming scientific discoveries into practical technologies.

Yet proposed changes to federal research policy could make this system less predictable. For instance, one agency has proposed revisions to its financial assistance guidance that would give the agency a greater role in licensing patented inventions that benefit from its funding, rather than leaving those decisions to the institutions, as under Bayh-Dole. And a detailee at a federal research agency went even further, proposing that the Bayh-Dole framework be abandoned altogether. 

Changes like these may sound procedural. But successful technology transfer depends on predictable rules. Researchers, universities, investors, and companies need confidence that the terms governing funding and intellectual property will remain stable throughout the long process of bringing an invention to market.

Before policymakers change those rules, they should be confident they aren’t obstructing early-stage discoveries from becoming real-world products.

When my team and I developed the camera-on-a-chip, none of us could have predicted all the ways it would eventually be used. That’s often true of breakthrough innovations. And one of the Bayh-Dole Act’s greatest strengths is that it gives discoveries the opportunity to surprise us. 

Policymakers would do well to protect that system proven to work, instead of trying to fix something that isn’t broken.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.