Going Beyond The “Buzzword”: Bringing the Future Potential of AI into Better Focus

If artificial intelligence is going to have an overall beneficial presence for tomorrow’s medtech devices, today’s industry leaders will need a better understanding of how it needs to evolve.

Joe Darrah

October 18, 2024

9 Min Read
Image created in Canva

At a Glance

  • AI remains a buzzword in healthcare, often misunderstood in its application to medical devices.
  • As AI technology evolves rapidly, manufacturers face challenges in meeting quality standards and ensuring compliance.
  • Kellen Giroux predicts that AI will significantly impact both pre-market and post-market processes in med-tech.

Although the roots of artificial intelligence (AI) can be traced back through several decades and various iterations, its presence as a consistent part of everyday life is something that society is still learning to embrace.

When it comes to its role in healthcare, AI remains somewhat of a “buzzword” today, with its true meaning and implications often elusive, especially related to medical devices.

As AI’s capabilities expand beyond a mere feature of sophisticated technologies, these tools are reshaping quality standards across the board, from pre-market assessment to post-market surveillance.

As technology becomes more sophisticated, AI and machine learning will continue to transform diagnostic capabilities, predictive analysis, and overall med-tech innovation.

Impacts to safety and compliance will continue to deepen, and ethical implications will become more common. As such, it will be increasingly important for industry professionals to be more cognizant of how AI operates, how it differs from machine learning and generative AI, how it influences patient care, and how public perceptions compare to AI’s actual capabilities.

During his upcoming presentation at MEDevice Silicon Valley titled “QualitAI: How AI & Machine Learning Fit Into The Quality Future,” Kellen Giroux, director of quality solutions at Network Partners Group, Warsaw, IN, will seek to empower today’s medical device company stakeholders on how the industry can collaborate to chart a more productive future by engaging in more open dialogue about AI’s impact on medical devices.

Related:A Practical Future of Developing AI for Medical Devices

He will also discuss potential misconceptions and pitfalls as well as the potential for transformative benefits.

Giroux recently participated in a Q&A with MD+DI to discuss how various members of the medtech industry will have to come together to learn how to best integrate AI safely and effectively into various processes, services, and products.

When considering that AI remains a buzzword even within this industry, do you think most professionals “understand” how AI functions - or do you believe there remains a certain level of uncertainty and confusion?

Giroux: “I think there is definitely still a level of uncertainty and confusion. That is one of the primary goals of our [conference] panel - to educate and enlighten professionals regarding what AI actually is and how it functions. What they may deal with on a daily basis and may not realize, and how AI is being integrated into the medical device space. There are so many different types of AI (including machine learning, generative AI, and reactive machines), so it is difficult for people to understand what each one is and how they all function.

Related:AI Regulations: A Priority for Medical Device Manufacturers

The saturation of the term AI and the vast number of different types has contributed to this confusion. Movies and Hollywood have also skewed how AI is understood. I have found the perception to be at both ends of the spectrum, where people either believe AI is a pipedream or that it could end the world tomorrow. The reality is that AI is somewhere in the middle (for now).”

Do you think the presence of AI will make it more challenging for manufacturers to adhere to quality standards?

Kellen Giroux: “Yes, absolutely. Quality standards are about ensuring medical devices are safe and effective. Validating, justifying, and qualifying AI to prove that they make devices safer and more effective is truly where the burden of the work lies. Because of the rapid evolution to the different types of AI, regulators are struggling to keep up as well. When the innovations begin to outpace the regulations, compliance becomes a major challenge, especially if [the regulations] don't yet exist. Again, I think confusion around AI clouds that issue. The misconceptions of what AI is capable of can influence how it is utilized and regulated. A lot of the AI in the life science space is based on machine learning, which requires good data to go in for the AI to put good data out. That inverse relationship has caused both manufacturers and regulators to simultaneously underestimate and overestimate AI.”

Related:Using AI to Help Medical Devices Help Themselves

Moving forward, do you see AI having more influence on medical devices during pre-market processes or post-market process?

Kellen Giroux: “In the near term, the influence will be highest on the pre-market process. Most medical device manufacturers who are investing in AI are working to integrate it into their products and to develop AI-enabled interfaces. Right now, that is seen mostly in assisting with customer workflow, allowing the AI to anticipate needs and to make the device operation easier for the user. Eventually, with greater leaps in the AI learning process, the devices will begin to branch into the diagnostic space. We're not there yet for full diagnoses, but that is where many companies are trying to go. Conversely, I can foresee a heavier investment in the post-market phase in the not-too-distant future. The benefits of utilizing AI and machine learning to analyze field data, safety information, and customer experiences will greatly benefit how we investigate patient events and beyond. There are many companies developing these types of programs right now that can generate useful reports in this area, with some even able to offer predictive information so that organizations can anticipate issues, compliance gaps, and recalls. This will eventually feed into making better devices by organizing and predicting the customer needs. In the end, it will be a cyclical benefit.”

AI is progressing at such a rapid pace that the term “innovative” seems to be constantly expanding. What is that you think is truly innovative about AI in the medical device space as we get closer to the conference?

Kellen Giroux: “I think as we see more innovation in the diagnostic space, we are going to start to see some truly remarkable products that are able to determine patient symptoms and outcomes. As I mentioned, I have not seen devices that are really living in this space yet, but, with the proper database, I don't think we are too far off from this really taking off. The current capability that I believe is truly innovative is the ability of some post-market AI. There has always been a plethora of data available, but analyzing it, understanding it, and utilizing it has been difficult for experts and organizations to sift through. Some programs now can analyze public patient incidences and guide a company towards how best to risk assess the same hazard in their own devices, thus allowing organizations to be more proactive and to improve patient safety. I am working with such a company, and it is genuinely awesome to see in action!”

How has AI impacted the need for companies to invest in data and analytics? What is the significance of this type of investment in relation to the presence of AI in med-tech?

Kellen Giroux: “I have always been a huge advocate for data and analytics in the med-tech space. Not many companies truly invest in this area. I think the companies that do get a great benefit for their efforts. It allows you to see market trends and keep a finger on the pulse of how their devices are being utilized. It provides an amplification to the customer's voice. With the presence of AI in med-tech, these trends, signals, and data will be even more important to keep track of, otherwise the data feeding the AI brains will be incomplete. I can see companies investing more in data and analytics as AI in med-tech grows, especially if AI in data analytics continues to grow commensurately.”

Are there any common misconceptions about AI that you plan to address during your presentation that can be previewed here?

Kellen Giroux: “One of the most common misconceptions revolves around AI being objective and unbiased. With the current iterations of most AI, they are only as good as the data being fed to them. If the data has a lean to it, the AI output will as well. The ‘reasoning’ of AI is still very limited in that regard. For example, AI had recently been utilized by certain court systems to suggest sentencing recommendations. A study by Tulane University found that the sentencing recommendations by the AI were found to be prejudicial towards certain groups of people, suggesting longer and harsher sentencing guidelines. It was concluded that this was because the historical sentencing data used to ‘train’ the AI was also prejudicial in the same manner. Until biases can be removed, or data can be baselined, this type of output is still very likely.”

Can you provide a few insights about the differences between AI, machine learning, and generative AI that you intend to discuss more at length during your presentation?

Kellen Giroux: “AI is the overarching concept that describes this vein of technology in general, and this is where a lot of misconceptions come from. To some, AI can mean ‘artificial intelligence,’ aka a sentient being. To others, AI can mean a high-tech calculator. It depends on your experience. Machine learning is one of the most common forms of AI that people have interacted with. It focuses on learning from previous data. Generative AI takes that a step further and uses its data to create new things, rather than just a report or combination of the data it already has. We will be covering examples of these and how they are being utilized in med-tech [during the presentation].”

When you think about the future of AI in medical devices, what do you see as an example of the greatest benefits to manufacturers, providers, and patients. What are the most significant challenges?

Kellen Giroux: “Manufacturers will see significant benefits from AI in both the development of their products as well as in the products themselves. They'll be able to offer enhanced patient experiences and use AI to analyze them to create better products. Patients will see an even more comprehensive experience with their provider, who will be able to offer faster results, greater information, and cover more difficult procedures in the near future. As with most rapidly evolving technologies, the challenge for all will be when and how to utilize this powerful tool. Reliance on AI to perform things could result in negligent outcomes if these tools are implemented before they're ready or before the surrounding environment can adapt. Caution and strong risk assessments will be crucial in integrating this technology into the industry in both a safe and effective manner. In either case, AI in the med-tech space is inevitable, so guidance in this area will need to be collaborative as we all learn together the best paths forward.”

About the Author

Joe Darrah

Joe Darrah is an award-winning freelance journalist based in the Philadelphia region who covers a variety of topics, including healthcare and medical technology. His articles have been published in more than 40 publications.

Sign up for the QMED & MD+DI Daily newsletter.

You May Also Like