DeepHealth的动态

Transcript

Well, thank you so much for making the time, Professor Rockwall. Clearly at the CR, AI is in everybody's mind. What are some of the challenges and trends you're hearing about AI? For sure, at the moment, I think AI deployment is the big issue that radiologists are facing today. We've seen huge developments with the companies getting AI products across FDA approval or CE marking, and now we have to get them across the next hurdle and into a delivery of. Health care and I think we have lots of different challenges, some of them are technical, such as interoperability and some of them are human, such as what am I going to, how am I going to work with this new AI tool? And there's all the concern about trust and about bias. So we have some different challenges with deployment and implementation, which I think we're working through at the moment in our community. That very much resonates with me is, by the way, one of the missions that Deepal definitely. Into sulfur and it's so important because you can think about the importance of AI in population health. It clearly has an impact. So for instance, we think about targeted lung health check program in the UK that we're partnering up with with the NHS that's detecting cancers and a much earlier stage than without AI. He clearly does have an impact on patient outcomes as well as the post OK, So what are you what are your thoughts on driving AI further as part of population health screening program? I mean I've always seen. Breast screening as one of the early adoption areas obviously for AI where in the NHS as you mentioned, we do double reading for breast AI and we have an enormous workforce, shortage of big workforce probably not just in the UK but really globally. So when we want to look at population screening to have the assistance of AI to help with with this mission if you like is tremendous because I don't think we even have in our training program. Enough space to train, enough to really branch out to screen for breast, lung, Ideally prostate will be coming. So the AI tools are going to make that possible, working together in partnership with radiologists. I can't wait to see that implemented. I think it is going to drive early detection, which is what we need in healthcare. Early detection, by the way, is also very sustainable because patients are picked up early and they need much less complex care on the whole. Than patients that are picked up later on in the stage of the disease. So it's not only beneficial for the patient's health, but it's also beneficial for planetary health double impact as we say and, and and you talk actually about the role of the radiologist as part of that as well and in what ways that evolving. Certainly at the moment where I work it's not legal to have fully automated AI in diagnostics. Now of course there are other AI tools. Workflow area, which can obviously go without a radiologist in the loop, but when it comes to diagnostics, by law there has to be a final sign off by a radiologist. So there's quite a journey that we're going to take now when radiologists are starting to use AI, learn to trust AI where appropriate, and also to learn where the pitfalls are. And I think radiologists need to be an integral part of finding this out and studying this whole area of the human AI. Interaction, I have no doubt in time there will be narrow AI areas where automated AI is going to be hugely beneficial. You touched upon many things. Just Eric, teamwork, we're only exciting journey and for sure ultimately AI will advance our mission. I want to thank you for your time, Professor Rockwell. It's lovely to meet you. Kate, likewise.