The discussion of government involvement in the light-speed adoption of artificial intelligence (AI) has cast a brighter light on the conundrum between unbridled product innovation and risk—whether real or perceived. With AI expanding in virtually every sector of the U.S. economy, the time to establish specific oversight and regulation is now. The key questions are why such oversight is necessary and how it can be effectively implemented.
AI has become a generic term that includes the ultimate use of technology to discover, optimize, and accelerate everything under the sun faster and better than humans alone. Sounds fantastic and inevitable, right? The danger is that, unchecked, AI could produce unnecessary risks in the areas of personal health, individual privacy, fair competition, national economic security, and other areas. The government can and should establish oversight to ensure that companies use AI responsibly to reduce risks to individuals and our economy.
Drawing parallels with the FDA's role in regulating medications demonstrates the benefits of combining government regulation with the advances of AI. Today, drugs take 10-12 years to get from discovery to patients and over 90 percent fail to make it through that process. Because of the long and risky development process, less than 50 new drugs make it to market each year. There are nearly 7,000 diseases of which we only have about 500 approved treatments. So, we have government oversight and regulations that are working to keep medication safe, but are limited by the volume of qualified drug candidates.
Using AI, companies are reducing the time and costs associated with the discovery and testing of new drugs. It's expected that AI-assisted drug development and testing will greatly increase the number of medicines. This is likely true for medical devices, testing, preventive care, and more. Without government regulation, however; the promise of more and better could be reduced to more and worse.
However, even with the adoption of AI, the FDA is still in control of the final product—to ensure safety as well as efficacy of the drug. In December 2022, the FDA Modernization Act 2.0 passed which clarified that pharmaceutical companies can use AI machine learning as an alternative to animal testing for non-clinical tests before a drug candidate goes into human trials.
In the near future, I expect the FDA will integrate AI into human clinical trials to increase safety and also decrease the time for approval. The result would be a larger number of safer drugs available to treat more diseases—including those that affect underserved populations. A government regulatory body with oversight over both the AI process and its resulting product will provide better outcomes compared to the stand-alone use of AI treatment. It also creates trust in the process and result.
Another scenario would be government oversight of data associated with AI systems. Using the health industry again as an example, sensitive health care data—including individual and population health data—is vital to U.S. economic interests. We are learning more and more that our health data is a national asset. In addition to uncovering health risks, U.S. companies use health data to discover and develop some of the most advanced health technology and methodology in the world. Using AI to understand how genetics and other biological data can predict health risks, prevent diseases, and extend life span is becoming the next big wave of the health economy.
Government oversight, including visibility into foreign ownership or control of U.S. companies using AI with sensitive data, significantly diminishes the risk of this information falling into the hands of foreign adversaries. This includes adverse impact on the health and privacy of individuals or groups through vigilant management of U.S. health care data.
It's important for everyone to benefit from these discoveries, and the U.S. economy should benefit from the development of new health discoveries from our own data. We need additional government oversight to keep U.S. companies from knowingly or unknowingly "sharing" sensitive health data with other countries that could use it for harm.
As AI evolves, the inherent lack of visibility into its decision-making processes may lead to a growing mistrust in the reliability of its results. Implementing regulations that include periodic audits can improve visibility of what AI is doing. While the "black box" inherent in AI's processes will likely continue, monitoring the functionality and implementation of AI systems can provide essential checks on potential unforeseen consequences.
While complete nationalization of AI is impractical, government oversight with audits and checks on AI systems and results will ensure the benefits of AI can be adopted without jeopardizing business interests or individual prosperity.
Tyrone Lam is chief operating officer of GATC Health.
The views expressed in this article are the writer's own.
Uncommon Knowledge
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
About the writer
To read how Newsweek uses AI as a newsroom tool, Click here.
