The U.S. Food and Drug Administration (FDA) recently announced that it will immediately introduce artificial intelligence (AI) technology in all its centers to shorten the time required for drug approvals. This decision was made after the FDA completed a generative AI pilot program for scientific reviewers. The FDA stated that with this AI tool, scientists can reduce a large amount of repetitive and monotonous work, thereby improving review efficiency.
Image source note: Image generated by AI, image authorization service provider MidJourney
Jinzhong Liu, deputy director of FDA's Center for Drug Evaluation and Research (CDER), said: "This disruptive technology allows me to complete scientific review tasks that once took three days in just a few minutes." Martin Makary, commissioner of the FDA, also mentioned that all centers will achieve full AI integration before June 30, a timeline that has drawn significant attention from the public.
The day before the announcement, Wired reported that the FDA was discussing AI applications with OpenAI, although the FDA's statement did not mention OpenAI or its potential involvement. Sources from Wired said that teams from the ChatGPT developer have held multiple meetings over the past few weeks with the FDA and two individuals from Elon Musk's government efficiency department to discuss the project named "cderGPT," which may be related to the FDA's CDER department.
Wired also mentioned that the FDA sponsored a study on large language models for internal use in 2023. Robert Califf, former commissioner of the FDA, said that the FDA's review team has actually been experimenting with AI technology for many years. "What is worth noting is the specific parts of the review where AI assists and its significance," Califf said. He pointed out that the desire to shorten the review time has always existed, and it is generally believed that AI can provide help.
In addition to the final approval review, the FDA is also considering applying AI technology to other business areas. Commissioner Makary expressed his enthusiasm for this technology, questioning why new drugs take more than ten years to come to market on Twitter and calling for modernization reforms using technologies like AI.
However, the risks of using AI technology in the medical field are also a cause for concern. A former employee who previously worked at the FDA mentioned that tests of ChatGPT in clinical tools showed that it tends to produce misleading information, a problem that will not disappear in the foreseeable future. As a result, there is widespread concern about the reliability of AI in reviewer tasks.
Key points:
🌟 The FDA is introducing AI technology to accelerate the drug approval process, with full integration expected before June 30.
🤖 The AI tool can significantly reduce the repetitive work of scientific reviewers and improve efficiency.
⚠️ Risks associated with using AI technology in the medical field have raised concerns, with the possibility of misleading information.