In the fifth and final part of this video interview with ACT editor Andy Studna, Stephen Pyke, chief clinical data and digital officer, Parexel, highlights the importance of human oversight when using artificial intelligence to aid in decision making.
ACT: How important is it that a level of human intervention remains with the oversight of these artificial intelligence (AI) solutions?
Stephen Pyke: It begins with the observation, which I think we all have heard, again, as part of this sort of public discourse here, which is AI, at this time, makes mistakes; sometimes it makes mistakes. There's this notion of hallucinations, which I think many people will have heard about, even if they're not entirely sure what it means whereby the AI can produce as fact, something which is entirely made up. And it's to do with the way the technology works. But I think the key point here is, this is a technology which is not 100% reliable, and therefore the important role of experts remains as necessary today as it's always been. AI is therefore I think, best seen as enabling good decision making, as providing information and evidence, which ultimately will be judged by physicians and other experts. And the human will make the decision. So we talk about human-in-the-loop, not just as a kind of catchphrase, which sounds good, but actually is an absolutely integral, necessary part of the way that we think about AI. I think there are other sectors where we're hearing about amazing things being done autonomously by AI. AI is making decisions. That's not the space we're in. That's not the way we're going to operate. And I think there is unanimous agreement across not just life sciences, but also healthcare more generally, that we don't want to be operating in that way. That's not the role for AI. But I think we all agree on this, to see human decision makers supported by really powerful AI technologies, speeding up processes, helping us get to better decisions more rapidly, and more often. And then we will see where we get to; maybe there was a lot a little different in five years’ time, this technology is moving very fast. But I think here and today, I wouldn't want an AI to make a decision about my disease, my treatment, I still want and believe it's essential that we have a critical role in decision making for physicians, and for other experts. So yeah, it's great that you raised this human oversight. It’s absolutely critical that we recognize its important role.
Regulatory Compliance With eCOAs
April 26th 2024In the fourth and final part of this video interview with ACT editor Andy Studna, Melissa Mooney, director, eCOA sales engineering, IQVIA discusses how the regulatory stance on electronic clinical outcome assessments has changed over the years and what it could look like in the future.
Using Patient Reported Outcomes in Dermatology Trials
April 25th 2024In part 3 of this video interview with ACT editor Andy Studna, Melissa Mooney, director, eCOA sales engineering, IQVIA sheds light on the unique challenges of dermatology trials and how clinical outcome assessments can be implemented in them.