Request Copyright Permission

Artificial Intelligence Formulation Science for Solutions: Breathing New Life into Respiratory Drug Delivery with Self Driving Laboratories

Shorthouse D, Cook M, Abdalla Y, Murnane D.

Respiratory Drug Delivery 2025. Volume 1, 2025: 94-100.

Abstract:

Artificial intelligence/machine learning (AI/ML) guided formulation development is gaining significant interest due to its ability to accelerate medicines discovery, and reduce development costs and reliance on the expertise of individuals. Traditional formulation development is time-consuming and heavily reliant on empirical methods. Human formulators rely on heuristic knowledge to efficiently narrow down formulation options, select suitable excipients, and define processing parameters all while considering time, resource constraints, and the target product profile. AI/ML can utilize datasets of a priori information to optimize formulation compositions and manufacturing parameters, streamlining formulation to achieve better outcomes more rapidly. This is especially valuable in an industry that faces constant pressure to bring new therapies to market faster and more efficiently.

We have developed ‘self-driving laboratories’ that leverage a coupling of AI/ML and robotics to bridge the gap between traditional formulation approaches and modern high-throughput experimentation, enabling automated data interpretation and intelligent experiment design to transform formulation science. Robotic platforms can rapidly prepare and test hundreds of formulation variants, while AI algorithms analyze the results to identify trends and optimize compositions and processing parameters, feeding back experiments for robotics to execute. This integration allows for the rapid identification of promising formulations with desired physicochemical and performance characteristics, transforming the way that we explore formulation spaces. We are currently building these integrated workstreams for the development of liquid formulations applicable to respiratory drug delivery. Here we discuss our progress with self-driving workflows, alongside opportunities within the respiratory drug delivery field.