Last month, CTO Scott Thielman attended MD&M West in Anaheim. 

One of the keynotes, led by Jennifer Samproni, Chief Technology Officer of Health Solutions, highlighted the significant impact human-machine interfaces (HMI) will have on the future of medical device innovation. Samproni elaborated on strategies for medtech enterprises to navigate product development and ensure market competitiveness.

AI, machine learning, IoT, and automation are increasingly important in healthcare. Research suggests these technologies have the potential to make healthcare more efficient and productive, which can enhance the capabilities of medical professionals and improve healthcare overall.

However, there are many aspects of these applications that remain questionable.

In her talk, Jennifer emphasized that medtech companies must embrace design principles that support patient-centric care and intuitive user experiences. 

One fascinating part of the presentation was the fireside chat with Adam Scraba, Director of Product Marketing at NVIDIA, on the intersectionality of HMI, AI, and machine learning. Below is a summary of the key takeaways from their conversation. 

AI Origins and the Future of Generative AI

While AI has traditionally been associated with tasks requiring logic and pattern recognition, generative AI goes further by simulating human-like creative processes, including empathy. 

Here, Adam highlights the evolution of generative AI and its origins. 

ImageNet is a famous benchmark for when we first had a machine predict what was in an image better than a human could. It was clinically proven that we now had superhuman vision, and that was only ten years ago. 

“Empathy will be built into our products. We already have products with AI that can understand context. Now, we’re beginning to see products we can talk to, and talking to an application means it understands our language and intentions.”

What questions can an application with built-in empathy answer?

• How would you like to interact with the application?

• What tasks should the data perform?

• What's the purpose behind clicking a button?

• What is the intention of this product, question, or action? 

For example, it can be used in:

• A manufacturing setting doing defect detection

• Education and training environments for interactive learning 

• Supply chain management to optimize inventory, demand forecasting, and logistics planning

• Personalized medicine by analyzing patient data, medical literature, and genetic information to support diagnosis and treatment planning

• Environmental monitoring to detect pollution or predict natural disasters

Here, Adam likened the future of generative AI in healthcare to the 2002 Tom Cruise blockbuster “Minority Report” (set in the year 2054) and the futuristic prediction system built on the application of our networks, products, and services. 

Empathy, he believes, will unlock a lot of amazing capabilities in healthcare. 

Here’s another example Jennifer and Adam shared. 

Historically, transitioning from one provider to another proved to be a complicated experience for many, especially those with a deep medical history. 

Jennifer recounted her recent experience trying to switch providers, which involved transferring her results from diagnostic scans and screenings. Her prospective new provider recommended returning to her previous one due to the time-consuming process of reviewing and condensing her records, which frustrated and disheartened her.

“A real human element is needed to review minor changes in a patient’s healthcare history and results year over year. AI's remarkable predictive and diagnostic capabilities and extensive healthcare datasets can take that predictive power in healthcare to the next level.” 

Generative AI can distill all these data points into optimized, salient, and effective information we can implement. Manually doing this is expensive and time-consuming. This technology truly is transformative

Digital Twins and Surgical Robots

Another exciting application of generative AI in healthcare is the creation and use of digital twins and surgical robots. 

Digital twins enhance surgical training by enabling personalized treatment simulations and predictive modeling. Utilizing 3D models and augmented reality, surgeons can also practice procedures virtually before performing them in real time. 

On the other hand, surgical robots can provide immediate feedback during the procedure and optimize the process for better patient outcomes. 

Adam explained that NVIDIA, initially recognized for its gaming platform, has expanded its focus to include simulation, which is really just a form of accelerated gaming. This evolution has led to the developing of a business centered around "digital twins."

“Our AI will be trained, tested, and enhanced in the digital world more than a person could ever be in the real world. Surgeons can now train in simulation. Just as they train autonomous vehicles to drive, surgeons will inevitably be trained using augmented intelligence.”

What does the future look like?

So, what does the future look like? How can AI, machine learning, and large language models be integrated? What does that look like for medical devices? 

Adam shared the themes and trends he expects to see in the coming years:

• We will be interacting/communicating with human-machine learning on a daily basis. AI agents (not chatbots) will be buried in our software and hardware. 

• A software-defined nature (or technology approach) to healthcare will further unlock precision medicine.

• Outside the hospital, generative AI will be used to discover, generate, and test potential drug molecules.

All these intersection points are rapidly expanding, creating an exciting future for innovation in medical devices.