The Support of Artificial Intelligence in Biologics Production Volume: A Conversation with Alessandro Zannini, Global After-Sales Manager, Stevanato Group

PE: Pharmaceutical companies have been increasingly relying on artificial intelligence to increase production volume for biologics. Can you briefly explain this process, and what makes it so effective?

Zannini: When you look at the production of biologics and specifically in our field of inspecting glass containers, utilization of AI is allowing the company to focus more on the quality of their products, also getting a higher quality product out for patients. This is because AI is used in specific tools to help identify the most critical areas of the application with that system. Therefore, it is absolutely possible to discriminate much better from what is a true defect and what is not. It avoids costly re-inspection and more in general with the data collection, it's also possible to improve on that.

PE: How exactly does AI contribute when it comes to keeping supply chains organized and how does it help to prevent potential disruptions?

Zannini: This isn't in our field of application, as we don't work much with supply chains in terms of application AI. However, even if you still compare it to being able to minimize the waste and unusable products, it will facilitate the supply chain allocating some of the needs that are in that field.

PE: In its current state, do you believe that there’s potential for generative AI to make substantial mistakes regarding the development of biologics?

Zannini: We work i a very controlled environment that is FDA approved. The use of generative AI is not yet implementable in the area where we are. Basically, the utilization of the AI that we have is used before the inspection. What this means is that while in the future, there might be some development in this area but right now, it is something that is used upstream from where we come into play with the AI.

PE: As artificial intelligence continues to develop, where do you believe improvements can be made regarding production volume in the next decade?

Zannini: We are at the beginning of utilization for AI in many fields. I'll give you a brief explanation of how it is done for us. We use AI to really zoom in and rebuild the tool that is then inserted inside of our equipment that is AI generated, and that tool is frozen. The complexity of that tool is limited, because when we inspect containers at the rate of 600 per minute, we are talking milliseconds of time to do the elaboration. Where the technology is going to help us is in developing more powerful miniaturized computers that will allow more complex models to be used. Therefore, AI will become more effective in the future.