The use of artificial intelligence (AI) is seeing a surge in a variety of industries. But does the regulated nature of bio/pharmaceutical development, manufacturing, and commercialization serve up unique challenges for incorporating AI into processes? According to Siegfried Schmitt, vice-president, Technical, at Parexel, using new technology will most likely require the assistance of experts in that technology. And if you’re contracting a third party for that help, you need to do an assessment of that vendor.
“[T]he thing with AI, it being a computer-based technology, there is sometimes the issue of, do you have sufficient or good documentation practices applied? It's all too easy to play around with computer code and settings and so on, but you do need to have appropriate documentation, and because you're applying it to process development … It is something that can get inspected, so be ready for an inspection,” Schmitt cautions.
Susan J. Schniepp, distinguished fellow at Regulatory Compliance Associates, stresses that a robust quality management system and document control are necessary when adding AI to a process. “I also worry about if there's the expertise to interpret what the data [are] telling you. You're going to be collecting a lot more data than you've collected before, and data that's new, and how it impacts your organization. And the data you're collecting may prompt you to want to make a change in the coding, [but] you need to follow that stream through your quality management system,” Schniepp says.
Watch Sue and Siegfried answer the question: “My company is researching the possibility of using AI in the formulation of a new product. Are there any quality considerations we should be aware of?”
Susan J. Schneipp is distinguished fellow at Regulatory Compliance Associates and a member of Pharmaceutical Technology®'s Editorial Advisory Board.
Siegfried Schmitt is vice president, Technical at Parexel and a member of Pharmaceutical Technology®'s Editorial Advisory Board.