Artificial intelligence (AI) will impact every part of the drug life cycle, including production—according to the European Medicines Agency (EMA), which says dialogue is vital for the development of effective regulations.

The European Union (EU) drug regulator made the comments in a concept paper published recently in which it urged drug manufacturers to take a human-centric approach to the development and deployment of AI and machine learning (ML). Growing industry interest in AI prompted the paper, according to EMA communications officer Anna-Sofia Joro.

“Our stakeholders are exploring the use of AI in many areas of drug development. Manufacturing is a structured, large-data, high-frequency activity, with potential for AI/ML automation. Process development and quality control are only two examples, which have been mentioned as the interest in the use of AI in these areas is growing,” she says.

“AI is an enabler for Pharma 4.0—the integration of digital technologies into pharmaceutical manufacturing. It offers the potential of increased process and product understanding and control, reducing development times and waste, enabling automation, real-time monitoring and control, facilitating analysis of trends, etc.”


In the paper, the EMA predicts use of machine learning—a branch of AI in which data is used to “train” computer models—will increase, citing process design, scaleup, quality control, and batch release as likely areas of application. And with this in mind, the agency says firms that use AI tools to create digital versions of production processes will have a lot to consider. “Model development, performance assessment, and life-cycle management should follow the quality risk management principles, taking patient safety, data integrity, and product quality into account.”

It also urges developers to think carefully about data used to train models.

“AI/ML models are intrinsically data-driven, as they extract their weights from training data. This makes them vulnerable to the integration of human bias into models. All efforts should be made to acquire a balanced training dataset.

In addition, the EMA says developers should follow principles set out by the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) in Q8, Q9, and Q10.

The Amsterdam-based agency also expects AI to play a key role in the production of “personalized” patient-specific medicines, on the basis that it will allow manufacturers to tailor products to individual needs more efficiently.

“AI/ML can be used to individualize treatment in relation to factors such as disease characteristics, patient genotype, wide-band biomarker panels, and clinical parameters. This could include patient selection, dosing, de novo design of product variants, and selection from a pre-manufactured library of variants.”

Previous articleThermo Fisher’s Framework De-Risks Cell Processing Equipment Selection
Next articleCrucial Tips for Bioprocess Transfer