Since the public debut of OpenAI’s GPT-4 little over a year ago, artificial intelligence (AI) and machine learning (ML) have been the subject of endless, and often breathless, media coverage and analysis — a source of both anxiety and elation over their potential to transform the way we live and work.
In the life sciences, these technologies have potential (and actual) applications across the innovation cycle, from discovery through to post-market monitoring. In the spirit of sifting through the hyperbole and identifying the real potential, Clarivate convened a panel discussion entitled Let’s get real about data and AI. Here are some key takeaways:
- AI and machine learning will have applications throughout every phase and aspect of drug development and will open up previously unreachable frontiers of medicine.
- “AI is going to drive us into areas we have traditionally avoided because of the level of difficulty like researching rare diseases or genetic differences,” said Anne Marie Finley, President of the Biotech Policy Group. “As a result, almost 95% of rare diseases have no therapy whatsoever and frankly, at the moment, no hope. That’s going to change radically with these tools. I think it’s going to alleviate a lot of patient suffering and we’re going to see light years of progress in the next decade or two in areas where we had no hope at all.”
- The rapidity of AI and machine learning advancements is overrunning regulatory purview. Agencies are hustling to catch up, and the industry is struggling to move forward with these technologies in the absence of safe harbors. Payers, too, will be challenged by AI- and machine learning-enabled advances in medicine.
- “Because of this speed and the rate of learning, governments won’t have enough money [to pay for all the new therapeutics it produces],” said Darrin Baines, Global Head of Health Economics, Life Sciences Consulting Services at Clarivate. “For drug companies, there’s going to be more pressure on them to demonstrate value and demonstrate priorities. AI advances could make a whole range of medicines obsolete, and the new medicines will be really expensive. So, we’re going to have all these challenges around how governments set priorities and how industry lobbies and provides evidence”
- At Clarivate, we are taking a proactive approach to help our partners stay ahead of emerging regulatory frameworks like the E.U.’s Artificial Intelligence Act and the U.S. FDA’s Good Machine Learning Practice principles. For example, we have developed a thorough risk assessment and classification document to evidence the formal classification of the solution vis a vis established acceptable and non-acceptable use cases, and to detail how we will be undertaking compliance to in-scope aspects of these regulations.
- These technologies will suffer from blind spots caused by a lack of diversity as long as the data and the human experts feeding and managing them do.
- “These are machines,” said Dee Chaudhary, Principal, Life Sciences Consulting at Clarivate. “They have no ethics. They’re wonderful for compiling data but they have no ethics. Human beings that have worked in this field are going to have to be able to establish guidelines and ethical standards and to curate these outputs in a manner that includes things like diverse populations. Machine learning cannot do that of its own accord.”
- Cultural and experiential differences between tech and life sciences workers are an impediment to further development of these tools. Both groups can have a poor appreciation for the value of understanding the patient experience and the importance of UX in designing products.
- “There is a tech arrogance,” said Omar Manjewala, CEO of Dario Health, “an assuredness that we understand your problem and we can solve it. I’ve seen a lot of companies fail because they did not understand the problem. They assume that the problems in healthcare can be solved with a simple solution without really understanding the entrenched forces at play that are designed to create market inefficiencies, the sort of counter-intuitive forces at play in the healthcare ecosystem.”
View the whole conversation here, and you can learn more about how Clarivate helps customers make critical decisions, navigate roadblocks and achieve their full potential using AI and machine learning here.
