[Adobe Stock]
When asked to assess the potential impact of the center, Charles Fisher, founder and CEO of Unlearn, said: “The truth is, I don’t know. I’m sort of in a wait and see kind of approach to understand what impact this new initiative is really going to have.”
One reason for the reservation boils down to resources. “I think that the agency is trying to do the right thing, but I think that these programs are dramatically under-resourced,” Fisher said. Getting significantly more funding could require a Congressional appropriation or new legislation to allocate additional resources to the agency and its programs. As things stand now, the agency has requested additional funding for FY 2025. In all, it is seeking $7.2 billion in annual funding, an increase of $157 million over the prior year. A sliver of that amount is focused on supporting IT and infrastructure
While AI is not an explicit focus of the C3TI initiative, the FDA has increasingly integrated machine learning (ML) into various aspects of drug development and review processes. For instance, the Center for Drug Evaluation and Research (CDER) has more than 20 different AI use cases underway. Examples include the use of a natural language processing system to partly automate the categorization of adverse event reports and using ML for data augmentation to detect anomalies in pharmacokinetic data for generic drug reviews. Another AI project is designed to streamline nonclinical study report analysis while the Information Visualization Platform (InfoViP) intends to support decision-making through data visualizations.
A talent war makes skilled workers scarce
Yet the FDA, like the broader life sciences industry, must grapple with a critical shortage of skilled AI professionals. AI has been much ballyhooed for its potential to drive efficiency across industries, but mature AI programs demand mature talent, and that is in short supply. Consider that several tech firms are dangling million-dollar compensation packages to lure potential AI candidates. Also in tech, some CEOs are getting personally involved in recruiting efforts, reaching out directly to prospective hires with pitches about opportunities at their companies. Meanwhile, many pharma companies are deploying AI programs on an “ad hoc” basis, as Deloitte has noted.
Charles Fisher, Ph.D.
The FDA’s Center for Drug Evaluation and Research (CDER) formed an AI steering committee in 2020 to track AI-related activities, identify best practices, and promote appropriate use of AI within the center. While the FDA is well placed to find experts in fields like biostatistics, it may prove more difficult to find seasoned AI experts. But the life science sector is generally struggling to find top-tier AI candidates. “There’s a lot of competition,” Fisher explained. And while some tech companies can justify spending millions on a senior data scientist salary, “no pharma company is doing that,” he said.
While there are exceptions, he noted, “for the most part, it’s very difficult to bring AI talent to the entire life sciences, pharma biotech industry in general, let alone to the government agencies.”
Misconceptions and cultural hurdles in AI adoption
Another barrier to adoption of AI in the life sciences industry is confusion about the nature of AI and entrenched cultural barriers. While interest in AI in healthcare has surged recently, Brookings noted in 2022 that there had been “little AI adoption in health care.”
Fisher notes that a common misconception in life sciences is a reductionist belief that AI is a form of statistics. But while statistics and mathematical models are core underpinnings of neural networks and other ML algorithms, AI is fundamentally a software engineering discipline, Fisher noted. Yet a cultural divide between software engineering and life sciences communities often leads to AI professionals feeling undervalued, resulting in an “organ rejection” phenomenon that could potentially leave some AI experts sidelined. “It’s really a people-related problem, not a technology-related problem,” Fisher said. “That has to do with the culture of the industry.”
A classic innovator’s dilemma
The life sciences industry is facing a classic innovator’s dilemma when it comes to the adoption of AI technologies, a concept popularized in Clayton Christensen’s influential book, “The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail” (1997). Christensen argues that successful companies often fail to adopt disruptive technologies because they are focused on serving their existing customers and maintaining their current business models, leaving them vulnerable to incumbents who embrace disruptive technology.
The rapid pace of AI advances means that companies that fail to adopt these technologies today may find themselves at a significant disadvantage in the near future. Fisher noted, “If that’s true and you missed the boat today, five years or 10 years from now, you’re gonna be really unhappy that you didn’t try it, but you didn’t adopt these technologies, right?”
Fisher argues that the root of the problem lies not with regulators, but with the industry itself. “The fact of the matter is that the pharma industry is less innovative in my experience than the regulators themselves,” he says. “We talked about Eroom’s law, right? The price and cost of drugs is increasing exponentially over time. Well, why is that happening? The truth is that people are causing Eroom’s law. It’s really the big large pharmaceutical companies themselves who are dragging their feet on innovation. It’s a cultural problem with the industry, not so much a regulatory problem.”
To overcome the innovator’s dilemma and fully harness the potential of AI, the life sciences industry must undergo a cultural shift that embraces innovation and collaboration with the tech community. Only by bridging this divide and investing in AI talent and infrastructure can the industry hope to keep pace with the rapid advances in AI and avoid being left behind by more agile competitors.
Filed Under: Drug Discovery and Development, machine learning and AI, Regulatory affairs
