FDA headquarters are located in Silver Spring, Maryland. (Credit: Ferdous Al-Faruque)
The U.S. Food and Drug Administration (FDA) has issued a warning letter to Prorea Cosmetic Research Institute for several violations of Good Manufacturing Practice (GMP). Notably, the letter warns the company of relying too heavily on artificial intelligence (AI) to create drug specifications, procedures, and production records without proper quality control oversight.
lawyer who listened concentration It noted that this measure does not mean that the authorities are against AI. Rather, it emphasizes that manufacturers should not rely on AI without properly understanding GMP.
The warning letter to the Livonia, Michigan-based company was filed after a three-day inspection in late October. The company manufactures homeopathic medicine products such as Dermveda Extra Strength Shingles Relief and Dermveda Extra Strength Ultra Genital Herpes Relief.
The overriding theme of the warning letter was that the company lacked a basic understanding of GMP and relied on AI to replace this knowledge.
The FDA said the company’s quality control department does not ensure that the drug is GMP compliant. For example, the quality control department under 21 CFR 211.22 failed to ensure that established procedures were followed, that batch records were reviewed prior to product release, and that appropriate production and process controls were in place.
Additionally, the FDA cited the company’s quality control department under 21 CFR 21.22(c) for “improper use” of AI in drug manufacturing. Company officials told the FDA that the company leveraged AI to comply with FDA regulations, including creating drug specifications, procedures, and master administrative records, among other things.
This did not appease the FDA. If AI is used to generate documents, companies “must review AI-generated documents to ensure they are accurate and in fact CGMP compliant; failure to do so violates 21 CFR 211.22(c),” the agency said.
Another issue was the company’s over-reliance on AI. For example, one investigator noted that the company did not perform process validation to ensure processes were under control. The company responded that it was “not aware” of the legal requirement because the AI did not notify it that this verification was required.
There were also other problems such as unsanitary conditions at the facility. During the investigation, investigators observed the presence of insects, dirt, fallen leaves, and clutter in several areas of the facility.
Additionally, the company did not take sufficient measures to prevent contamination from both internal and external sources. For example, when the docking bay door opens, the manufacturing area is directly exposed to the outside environment.
The FDA said the company failed to conduct “appropriate laboratory testing” to ensure the product did not contain nuisance microorganisms.
The FDA noted that the company has stopped manufacturing the drug at the facility.
Callie Richardson, a lawyer at Hyman, Phelps & McNamara, said the letter highlighted fundamental flaws in GMP and should not indicate that the agency opposes the use of AI.
“One of the lessons I learned is that I don’t get the impression that there’s anything wrong with the FDA using AI for documentation. The problem with the FDA is that no one reviewed it and it was completely inaccurate. I don’t think they’re saying it’s not OK to use AI. I don’t think this is what the FDA is saying… It seems like the problem was very elementary. No one knows anything about GMP, and no one knows enough about it to know what the AI is saying.” That’s wrong. ”
Mr Richardson said manufacturers should ideally use a closed system and know what their inputs and raw materials are. “I think it’s interesting that the warning letter focused on the output, and that the companies used this to generate something that they didn’t even understand. The FDA didn’t mention the input. My guess is that they were using off-the-shelf tools and weren’t trying to create some kind of internal system.”
prorea warning letter
