As the FDA explores AI’s potential to revolutionize drug evaluation, the agency must navigate challenges and complexities to harness this technology for accelerating pharmaceutical development.
The Food and Drug Administration (FDA) has been meeting with OpenAI, the pioneer behind the popular chatbot ChatGPT, to discuss the potential use of artificial intelligence (AI) in the agency’s drug evaluation process. The discussions are part of a broader effort by the FDA to harness AI technology to speed up the approval process for new treatments.
The Benefits and Challenges of Using AI in Drug Evaluation
According to FDA Commissioner Marty Makary, the current drug review process takes over 10 years, which is far too long. Makary believes that AI can help modernize this process and bring it into the 21st century. The agency has already completed its first AI-assisted scientific review for a product, and discussions are underway about using AI to analyze large amounts of data.
Dr. Marty Makary is a renowned American surgeon, medical researcher, and public health expert.
He serves as the Director of the Center for Surgical Innovation at Johns Hopkins University School of Medicine.
Makary has made significant contributions to the field of healthcare, particularly in cancer surgery and patient safety.
He has published numerous studies on reducing surgical complications and improving hospital efficiency.
According to his research, medical errors are a leading cause of death in the United States, affecting over 250,000 patients annually.
However, there are also challenges associated with using AI in drug evaluation. One concern is the reliability of AI models, particularly those that generate information based on patterns learned from large datasets. An ex-FDA employee who has tested ChatGPT as a clinical tool raises questions about how robust this platform will be for reviewers’ tasks.
The Current Mechanisms and Future Directions
The FDA currently has several mechanisms in place to expedite the review process for promising drugs, including the fast track designation and breakthrough therapy designation. These designations are designed to bring innovative treatments to market more quickly, but they also come with strict requirements that must be met before a drug can receive approval.
In terms of future directions, some experts believe that AI could be used to automate certain tasks related to the review process, such as checking for application completeness. This could help expedite the return of feedback to submitters and make the process more efficient.
The Role of Policy Guidance
Rafael Rosengarten, CEO of Genialis, a precision oncology company, emphasizes the need for policy guidance around what kind of data is used to train AI models and what kind of model performance is considered acceptable. ‘These machines are incredibly adept at learning information,’ he says, ‘but they have to be trained in a way so they’re learning what we want them to learn.‘
The FDA’s Research on AI
In December 2023, the FDA advertised a fellowship for a researcher to develop large language models for internal use. The agency is also exploring the potential of ChatGPT Gov, a self-hosted version of its chatbot designed to comply with government regulations.
As the FDA continues to explore the potential benefits and challenges of using AI in drug evaluation, it’s clear that this technology has the potential to revolutionize the way new treatments are approved. However, careful consideration must be given to the policy guidance and regulatory frameworks that will govern the use of AI in this critical process.
