P232: From Dreams to Implementation: The Realities of Incorporating GenAI in Life Sciences and Healthcare
Poster Presenter
Matt Wampole
Director, Solution Consulting
Clarivate United States
Objectives
The objective of this study is to provide thought leadership on how GenAI is impacting the digital health landscape for the life science and healthcare industry and the realities that must be addressed to implement this new technology.
Method
We reviewed literature, patents, specialist databases, and consulted experts to summarize the current thoughts on the companies and use cases of GenAI in life sciences and healthcare over the past 5 years. This included challenges that need to be addressed for GenAI to succeed.
Results
Preliminary results indicate that there are areas of interest for the implementation of GenAI into R&D processes, such as generating new drug assets for testing, accelerating clinical trials, and streamlining regulatory processes. These use cases for GenAI are posed to disrupt the current approaches to the development of new drugs. Approaches in the pre-clinical space have had the most success in the implementation of GenAI for generating novel assets. Using GenAI to summarize data is another area of need for commercial, regulatory, and clinical use cases, but there are several concerns coming up. As new approaches are brought into production, they will need to address challenges such as: data ownership, data privacy, data bias, reproducibility of results, and managing trust in the results.
Conclusion
Much like the current adoption of machine learning into all aspects of research and development, GenAI has potential to assist in generating new discoveries and accelerating processes to bring new drugs to the market. This technology can fundamentally change how insights and decisions are made across multiple areas of healthcare by accelerating decision making. This potential and need for disruption in the market is leading to strong growth in the digital health sector. For these new approaches to be adopted, several obstacles must be addressed. Some of these are shared across use cases, such as recognition of who owns the data used to train the models and ensuring the results generated are reliable. Other use cases will face specific challenges such as data privacy if using patient data or addressing data bias when incorporating incomplete datasets. None of these challenges are insurmountable but will require dedication and vigilance from all parties involved to implement GenAI for accelerating how the industry brings new drugs to patients.
