In a special edition recorded at the Dubai AI Festival, Below the Fold teams up with Date With Tech Podcast, Launch.fndry, and Trescon to explore the future of healthcare innovation. Parminder Bhatia, Chief AI Officer at GE Healthcare, shares how AI is transforming healthcare—from reducing MRI scan times by 83% to democratizing access to diagnostics. He dives deep into building trust in AI and why emotional intelligence is critical, even in a data-driven world.
🎙️ Topics:
⏱ Timestamps
00:00 Introduction to the Podcast and Guest
00:36 Transition from Amazon to GE Healthcare
02:07 AI Innovations in Healthcare
02:52 Future of AI in Medical Devices
04:20 Building Trust in AI Technologies
06:39 Listener Questions and Quickfire Round
08:28 Closing Remarks
[00:00:00] This is,~ uh,~ date with Tech podcast here at the Dubai AI Festival,~ um,~ in association with Below the Fold Launch Foundry and our wonderful partners recon we're joined today by Parminder Bhatia, who is the Chief AI Officer at GE Healthcare. ~Um, ~you are a AI leader,~ uh,~ especially in the,~ uh,~ med medical and, and ~uh, uh, ~healthcare space, and you've got an incredible,~ uh,~ background,~ um,~ which we'll touch upon in a moment.
~Um, ~we're gonna call today's session around precision and progress. Mm-hmm. How AI can really cut through in the world of healthcare. ~Um, ~so you led innovation at Amazon, I saw. Yep. And also now ge. How do you feel the, the shift has been going from an e-commerce giant and a and a cloud development giant going into the GE healthcare space, what has been the biggest change for you?
Yeah, so, ~uh, ~at Amazon I was leading healthcare for a substantial time. So I've been in healthcare last three to four years. ~Uh, ~at Amazon I was,~ uh,~ leading,~ uh,~ science aspects of,~ uh,~ foundation models, generative ai, so capability similar to chat [00:01:00] GPT, right? Which came from Amazon, like bedrock. So I was working on that.
But as we were working on building these technologies, it became clear that this technology was more tailor made for healthcare than anything else in the past. It is multimodal. It can understand. Yes. Large language model came in. Yeah. But it is language text, other things as well. And then that really motivated me to get back into healthcare and getting into GE Healthcare, which is more integrated.
Into the workflow where a lot of things that we built at Amazon as cloud providers, we would wanted companies like GE Healthcare to build on those things as well. So I thought this time it's the right time, where now it's becoming, AI is really commoditizing, it's becoming integrated into healthcare. So it made a right time to join GE Healthcare and kind of embark on that journey.
Sounds incredible. ~Um, ~what I'm keen to understand is, as the end user of the products, as the patient in a, in a medical environment. How does they, how does that person benefit versus, say, the hospital, obviously. 'cause for me, the enterprise use case is very clear. The automation, the digitization of [00:02:00] data, the ability to support somebody's,~ um,~ prognosis, but the end user, how do you feel they benefit?
So, I'll, I'll give an example,~ uh,~ of one of the,~ uh,~ AI technologies. So, GE Healthcare has,~ uh,~ highest number of a approved app authorizations, which are AI enabled. So one of the examples is Sonic dl. So. Your MRI scan for heart usually takes 30 to 90 minutes. Now we have built out this new technology, which can reduce the scan time by 83% leveraging ai.
Wow. Now, what does it mean? ~Uh, ~in the past, if it was hard to bring kids into an R machine, of course for 90 minutes, you can bring them out. ~Uh, ~patients with arrhythmia who could not sit or hold their breath. Now can actually get, MR. Scans can get better diagnosis because MR can unlock a lot more details, which was not possible in the past.
So you're talking about bringing better care, increasing the access, which was not possible in the past. Do you think then there is an evolution of this journey that one day we'll see having these sort of capabilities at, you know, [00:03:00] in-house within everybody's domain, do you think, do you see that world happening and envision envisioned as true?
Yes. So, and that's where a lot of automation is gonna help. ~Uh, ~across those different areas as well. ~Uh, ~you're gonna see a lot of handheld devices or small devices, which will make it easier. ~Uh, ~one example, like you have these ultrasound, which is actually handheld device. Yep. However, it requires a lot of expertise,~ uh,~ for someone to use an ultrasound to get a good quality,~ uh,~ ultrasound for a heart.
Let's say someone has a heart condition. Can I just use a scan? No, because it requires a lot of rigors. So one of the ways in which AI can be used as like a copilot or a navigation. Kind of an experience, can it provide like, hey, move a little left, move a little. Right now you have a really good diagnostic quality image.
Yes. Now bringing in those technologies, that's something which we,~ uh,~ built out. But that has a huge implication because now,~ uh,~ folks, even with less,~ uh,~ skills and technicality can use that and use that to get a really high quality image scan as well. So you're gonna see a lot more of that. A lot more of virtual and remote components are starting to happen [00:04:00] as well, of course.
So you're gonna see a lot of that happening, and that's where AI is. AI is best used when you don't talk about ai, and I think that's, that's, that's where, so both of the technologies that I talked about. We, we don't talk about them as like, yes, this is ai. It is solving a problem, it's helping on getting the guidance, it's helping on reducing the scan time.
But yes, it uses ai. You've touched on a really interesting point there, and I think it comes down to,~ uh,~ trust and what do you see as the biggest barrier of trust in terms of what you do as a, as a, as an innovator in your industry to help both get clinicians and patients to be part of your journey and have that faith?
Yeah, because obviously, you know, AI does have its downsides and it has, its ~uh, ~you know, there's debate around it and it's kind of the ethics side, the ethical side of things, but I just wanna get your thoughts on, on that, how to, how to build trust. Yeah. And I think you brought an initial question of like, how does it change for like a MedTech company versus a startup or a, a technology enabler who's building these technologies?
~Uh, ~one of the [00:05:00] things,~ uh,~ being part of GE Healthcare MedTech, you cannot be 80%. You cannot be 85%, of course you have to be because people's life is at stake. You have to be right every time. And also making sure the technology works as expected. So building an AI algorithm usually is 10 to 20% of our time, 80% of the time actually goes into multi-site validation.
We validate that technology at multiple sites, get feedback, trade on that before making it prime time ready in a lot of ways as well. So that increases trust because one of the things like ACON deal, which the technology was talking about for Mr. It has gone through 35 million plus scans. So it's not a technology which is just out there.
It has been used in real world with 35 million plus scans. Of course, that's happened across the globe. So you're talking about building technologies? Mm-hmm. Bringing everyone into the loop, bringing physicians into the loop, bringing,~ uh,~ patients into the loop so that they feel comfortable, and then kind of incorporating based on that, the feedback from them, it trading on that, that becomes useful.
The other area which is useful is. The new technology is coming out. [00:06:00] It's not black box. I think one of the key things for us to think through is like, it's not just, it's gonna give me, okay, patient has Alzheimer's or not. It goes into the details and provides explanation or reasoning that brings in more trust.
So it's still going to be augmenting or providing inputs through the physician who will be making the last call, but now he has more insights as to why that output came out as well. Of course. And I think fundamentally it comes down to providing an enablement of, ~um. ~You know, like you say, a combination of diagnostic tech capabilities and also, you know, being able to prognos like what, what's happened next and how to deliver on how, you know, ~um, uh, ~patients aftercare for a long, long time to come and keep coming back and trusted solutions.
I won't keep you too long 'cause I know that you are a very, very busy gentleman. ~Um, ~I just had a few more questions have actually come in from some of our listeners that I wanted to quick fire over to you. Mm-hmm. What's a myth in AI healthcare that you wish more people understood? So AI is not just at the output.
Like I think whenever we talk about ai, it's considered at the diagnosis healthcare data is multimodal. 97% of the data is still [00:07:00] not used, so all your clinical decisions are made on just 3%. AI can help you understand that 97% of the data, and that way physicians can make better informed decisions. So I think having that insight that it's not just replacing doctors, it's automating, augmenting, and really optimizing things which are not there today in a lot of cases as well.
Okay. Fascinating. If you could automate a daily task in your day-to-day life, what would it be? Calendars. Calendars, calendar. ~Uh, ~it has to be calendar. It's still not automated. We have seen a lot of solutions come out at Gen ai. We still see popups. It's not still at the best case. Okay. Okay. Maybe the Launch Foundry can help you with that.
~Um, ~what's, and finally, what's an innovation in health AI that you are, that you are really interested in, that you think is gonna change lives in the next 12 months? So we have heard a lot of,~ uh,~ innovations or discussions around agents, agent tech, ai, technology. Now it's becoming the next pillar of,~ uh,~ like discussions.
One area where agents can become really useful is to help you understand this multimodal data. Yeah. We,~ uh,~ last year we announced,~ uh,~ a project called IS Health [00:08:00] Companion, which. Uses this information to help in what we call as virtual tumor board. So usually when a patient has cancer, in extreme cases, a tumor board is set up where a lot of,~ uh,~ specialists, radiologists, oncologist, surgeon, come together, come out with what is the best treatment option.
Now a lot of those things can be streamlined in a lot of ways where physicians can get that insights even before having that meeting set up in cancer. Every day matters and how can you bring those things together is really gonna help revolutionize a lot of these things as well. It's been an absolutely incredible talking to you.
Thank you so much for giving us those insights today. Parminder Bhatia, thank you for joining us on,~ uh,~ date With Tech and hope to see you soon. Enjoy the festival. Thank you so much. Thank you so much. Thanks for having me. You're most welcome. Thank you so, so much.