By STEVEN ZECOLA
In its Strategy for Artificial Intelligence (V.3), the Department of Health and Human Services (“HHS”) acknowledges that: “For too long, our Department has been bogged down by bureaucracy and busy work.” HHS promises that it will accelerate artificial intelligence (“AI”) innovation, including “accelerating drug and biologic approvals at the FDA.”
History shows that well-intended but cumulative regulatory intervention – more so than scientific complexity – is the primary deterrent to rapid technological progress. If AI is subject to the typical pattern of regulatory creep, its potential to accelerate drug discovery and development will be significantly reduced. To avoid this outcome, HHS should develop a plan that is premised on a zero-based regulatory approach. That is, each new technology such as AI should start with a clean slate and only the minimum requirements deemed necessary to show effectiveness and safety should be applied in the approval process for that technology.
The Pace of Innovation
Medical innovation has lagged the pace in the other sectors of the economy. As Dr. Scott Podolsky of Harvard Medical School observed: “Medicine in 2020 is much closer to medicine in 1970 than medicine in 1970 was to medicine in 1920.” Podolsky points to breakthroughs such as antibiotics, antihypertensives, antidepressants, antipsychotics, and steroids that have not been met with same impact as innovations in the later 50 years.
Two explanations have been offered for this phenomenon: 1) the inherent complexity of biological processes; and 2) the regulatory approval process.
As a benchmark for comparison to the following case studies, the development of 4G communications spanned less than a decade, with discussions starting around 2001, technical specifications being released in 2004, and the first commercial networks launching in 2009.
Regulatory Intervention in New Technologies
- The Human Genome (Great Science Leads to Regulatory Paralysis)
The Human Genome Project (HGP) ran from 1990 to 2003, and has been lauded as one of the world’s greatest scientific achievements. The project identified the specific location of genes and DNA, creating a “roadmap” of the human genetic code and facilitating the identification of disease-related genes.
The HGP focused on balancing rapid scientific progress with ethical safeguards. Oversight was primarily managed through internal ethical programs and international data-sharing agreements rather than a single overarching legislative or regulatory body.
Under this structure, the HGP beat its target date by two years. That is to say that the complexity of the problem did not cause any delays, and progress was not impeded by the standard drug-approval bottleneck.
However, once the genetic roadmap was handed off for drug discovery and development, progress slowed dramatically.
After nine years, scientists were able to demonstrate that CRISPR-Cas9 technology was able to edit DNA. From there, it took eleven years for the first gene-editing treatment to be approved. After three more years had passed, the FDA announced that it was exercising greater regulatory flexibility for cell and gene therapies, noting that its processes have not always been fully clear to stakeholders.
Clarity has not been the only, or even the predominant, issue. CRISPR technology currently has countless applications, ranging from basic molecular research to advanced clinical, agricultural, and industrial uses. While only one CRISPR therapy (Casgevy) had been officially FDA-approved as of early 2025, there are hundreds of active research programs, clinical trials for diseases like cancer, and over 1,000 CRISPR screens for drug discovery. The regulatory resources to process all this activity is lacking. Further delay is inevitable.
In retrospect, we can see that once the HGP “roadmap” was handed off for drug discovery and development, the costs and the length of development ballooned under the regulatory review process. The delays seem to be attributable to a lack of FDA resources to administer its own regulations more so than the complexity of the scientific problem.
- Stem Cells (Regulatory Compliance Ignores Financial Constraints)
Geron began its human embryonic stem cell research program in the late 1990s, becoming a pioneer in the field and raising $100 million to conduct research and pursue blockbuster cures. At the FDA’s direction, Geron ran over 2000 experiments on mice and rats and exhausted a good portion of its funds in pre-clinical research. Geron’s Investigational New Drug application was comprised of 22,000 pages and cost $45 million. The company received approval to move forward with a Phase 1 clinical trial in 2009. There were no safety issues with the initial trial participants, but Geron terminated the program in 2011 for financial reasons.
Research on stem cells continue to this day for a variety of applications such as Parkinson’s disease and damaged optic nerves. In the meanwhile, Geron turned to oncology with its remaining funds and received the FDA’s approval in 2024 for a telomerase inhibitor after 34 years of research.
The Geron case is an example of what happens when regulation does not consider the time required to get to market or the cost of doing so. In other words, regulatory delay often indirectly retards innovation given the financial constraints on research. In this case, regulatory intervention also totally changed the direction of the research initiative.
- Personal Genetics (Regulatory Overreach Denies Innovative Medical Information)
Direct-to-consumer genetics provides an example where information itself became the regulated product. Founded in 2006, 23andme pioneered consumer access to genetic insights. In 2013, the FDA found the provision of personal genetic information to be a “device” that was “intended for use in the diagnosis of disease or other conditions or in the cure, mitigation, treatment, or prevention of disease” and that in certain circumstances “it could lead a patient to undergo prophylactic surgery, chemoprevention, intensive screening, or other morbidity inducing action”. Accordingly, the FDA found the provision of such information by 23andMe to be in violation of the Federal Food, Drug and Cosmetic Act by marketing its product without prior approval from the FDA.
The agency noted that 23andMe provided individual reports on 254 diseases and conditions and that the FDA expected 23andMe to get prior approvals for each of these reports. As the 23andMe CEO, Anne Wojcicki, explained in response, the FDA system would require over one million tests which would be practically impossible for either party to accommodate.
FDA granted 23andme its first approval in 2015, and continued to authorize various reports for direct consumer marketing over the next several years. However, regulation had altered the company’s direction and 23andme declared bankruptcy in March 2025.
In this case, the regulator stretched its reach into an information service for paternalistic reasons. That is, the FDA did not trust users with their own data. This had the effect of slowing down the pace of innovation and curtailing the value of the information.
These three cases show how years of scientific research can turn into decades of regulatory review and delay. The private and public equity markets are watching the regulation of AI in healthcare very closely to determine how much delay will be imposed on the first round of the AI-derived solutions.
The Current FDA Regulation of AI – Will History Repeat Itself?
As should be the case under the 1962 statute, the FDA requires every drug to meet a “reasonable assurance of safety and effectiveness” standard. Under the FDA procedures, AI-discovered drugs must also conform to the same conventional Phase I, II and III clinical trials for all drug candidates.
In addition, in January 2025, FDA introduced a 7-step framework for AI-enabled submissions. This framework requires sponsors to define the specific “context of use,” assess the risk level, and establish trust in the AI output before it is used for regulatory decisions.
The FDA also requires developers to provide detailed documentation on the algorithm’s architecture, training data, and potential biases, which is not required for traditional drug development.
In December 2025, HHS issued a Request for Information regarding the use of AI in patient care. As the initial comments in that proceeding establish, practitioners are concerned about liability in cases where they have used AI-technology in clinical care that the FDA subsequently finds to have required prior approval.
It is not clear which direction the HHS is headed as the market embraces AI applications in clinical care. However, given the nascent stage of AI innovation, HHS would have to pick the winners to accelerate the adoption of AI in clinical care. This would slow down the momentum in the market and add uncertainty. This uncertainty, by itself, would slow down innovation even without the imposition of additional regulation.
A Zero-Based Regulatory Approach
A zero-based regulatory approach would put an end to regulatory creep and focus regulators on the key requirements to demonstrate effectiveness and safety. Regulators would not participate in pre-clinical research. Documented evidence would be limited to addressing specific requirements to demonstrate safety and effectiveness.
The role of the government in such an approach would be as auditor to validate the output of an elongated trial. This function would include experimental validation, mechanistic understanding, and ethical oversight.
Such an approach would free up FDA resources so that many more trials could be run yet still receive the requisite oversight. With experience (and the application of AI), applicants and regulators could fine-tune the requirements as part of an annual zero-based regulatory review.
This approach is not deregulation. It is precision regulation. AI programs could make auditing faster, more relevant, and more accurate. There could be real-time data feeds to the FDA with automated auditing and exception reporting of key variables.
Conclusion
In an efficient market, the leaders would use AI to accelerate innovative solutions, streamline development, and deliver high-value experiences. The biggest threat to this outcome in the healthcare industry is regulation, not the complexity of the diseases.
The leaders of HHS should ask the staff to start with a zero-based regulatory scheme for AI, and impose regulatory requirements only where needed to demonstrate safety and effectiveness. This would not require legislative approval or an HHS rule change.
The role of the FDA would shift from being an overzealous gatekeeper at each step of the regulatory process to being a real-time auditor of scientific progress and the resulting innovative solutions. This approach would enable the staff to observe and assess many more innovative developments. We know that innovation breeds more innovation, which would have the effect of getting the healthcare industry back on track.
AI is at a nascent stage in healthcare. HHS should treat its regulations in that context. As it stands today, the current FDA regulatory framework will not accelerate AI innovation in the healthcare industry and should be adjusted before the initial timelines are missed.
Steven Zecola is a former technology executive and government official. He retired 24 years ago with a diagnosis of Parkinson’s disease. He currently is an ardent patient advocate.
Publisher: Source link
