TY  - JOUR
T1  - AI-Guided Bioprocess Control Should Optimize Critical Quality Attributes Before Productivity
A1  - Sven Larsson
A1  - Erik Johansson
A1  - Anna Nilsson
A1  - Lars Andersson
JF  - International Journal of Pharmaceutical Research and Allied Sciences
JO  - Int J Pharm Res Allied Sci
SN  - 2277-3657
Y1  - 2026
VL  - 15
IS  - 4
DO  - 10.51847/qVVVQ8OoyG
SP  - 24
EP  - 35
N2  - Artificial intelligence is increasingly capable of extracting process state from multivariate bioprocess data, forecasting trajectories, and proposing or executing control actions. Yet an optimization system can be technically sophisticated while pursuing the wrong objective. In biopharmaceutical manufacturing, maximizing titer, throughput, or equipment utilization does not by itself establish preservation of product quality. This Systems Perspective argues that AI-enabled control should therefore treat critical quality attributes as primary constraints or decision targets and productivity as an objective optimized within an acceptable quality-relevant operating domain. The distinction requires more than adding quality measurements to an existing productivity loop. Observed sensor signals, inferred biological states, predicted quality attributes, controller outputs, and validated product evidence have different evidentiary status and should remain separable. Multivariate process analytical technology, soft sensors, physics-informed and transfer-learning models, and adaptive control can shorten the path from observation to intervention, but each introduces uncertainty, transferability limits, and qualification requirements. We develop a quality-first control logic in which increasingly consequential actions require progressively stronger evidence that the relevant product-quality state remains within the validated domain. The resulting architecture is proposed as a systems-level manufacturing principle rather than an empirically validated universal control strategy.
UR  - https://ijpras.com/article/ai-guided-bioprocess-control-should-optimize-critical-quality-attributes-before-productivity-dz8cyfvdocz7jb0
ER  -