top of page

#007 - Challenges of Artificial Intelligence in Healthcare

There are three common challenges related to implementing AI in healthcare, and this is how they were in the news in the past week(s).

Navigating the challenges of AI in Healthcare
Healthusiasm Newsletter

Welcome to "It's a Healthusiasm World", a newsletter on Health, Business and Technology by Christophe Jauquet.

Every week (almost), I will amplify the positive changes that make our world a little healthier and happier.

If you haven't subscribed yet, join these 5000+ Healthusiasts on a mission to build a healthier future. Let's do it together.


The potential integration of generative AI in healthcare has sparked a mixed response from healthcare providers. While some are enthusiastic about its potential to revolutionise patient outcomes, increase efficiency, and uncover new insights, others take a wait-and-see attitude. Rapid technological advancements have indeed led to some early adopters of generative AI in healthcare. Yet, the speed at which AI is being introduced raises concerns about potential flaws and biases that might negatively impact health outcomes. 

Three common challenges relate to implementing AI in healthcare, and this newsletter covers how these challenges were in the news in the past week(s).

But first, this week's soundtrack to accompany this newsletter is a beautiful remake of a Groove Armada song. Indian folk band Midlake took on the electronic dance song "Roscoe - Beyond the Wizard's Sleeve" and put it in the perfect mood for reading. Quite the transformation as if an AI wizard pulled something from his sleeve.

Let's get the party started.

Software-related challenges: 

It can be challenging to know which software to integrate, when, and how to integrate it. The U.S. Food and Drug Administration has yet to approve any generative AI models among the 139 AI-related medical devices it has endorsed.

However, several notable initiatives have taken place: Academic institutions, like the University of Toronto, have embraced the potential of AI and established dedicated centres, such as the Temerty Centre for AI Research and Education in Medicine (T-CAIREM), to actively test AI innovation in healthcare. Meanwhile, Hugging Face and its partners have launched the Open Medical-LLM benchmark test to rigorously evaluate AI performance in healthcare tasks and ensure that these tools are reliable before reaching clinical settings.

What these recent initiatives show is, of course, the clear need for clarity. Which AI solutions can be used? I like how individual, valuable initiatives try to bring clarity while the larger healthcare systems themselves are still searching. (And there are way more initiatives than mentioned here, of course) However, whether this will have a sufficient or rapid effect in a centralised healthcare system is hard to predict. But at least we are gathering insights and making progress.

HCP-related challenges:

Despite this evolution, adopting "endorsed" AI solutions in healthcare will be as much more challenging. Many healthcare professionals lack the technical expertise to leverage AI fully, leading to a mismatch between expectations and what AI can achieve. On the other hand, AI specialists often do not have a deep understanding of healthcare, which can hinder the practical application of AI solutions as well.

The future of healthcare will likely demand a new breed of professionals skilled with a minimum of machine learning and health sciences. Some say that these ambidextrous professionals will be essential in leveraging AI to its full potential as the implementation of AI in healthcare evolves. But I wouldn't dare put it so firmly. There is a need for more knowledge, but I mainly see a greater need for trust in AI. I believe this can only happen if enough doctors are working on AI behind the scenes. So perhaps every doctor does not need to become ambidextrous, but some doctors need to be particularly committed to it. An essential question for tomorrow's physicians is: "Do we want to work in front of the scenes or behind the scenes?" And things are evolving in that direction.

In a recent episode of the Healthusiasm Podcast, we discussed how Imperial College London and Cardiff University are already offering (optional) innovation and entrepreneurship courses to medical students. (Click here to listen to the episode)

A podcast about trends and innovations in health and self-care
The Healthusiasm Podcast

Patient-related challenges:

Building trust among healthcare providers is crucial for implementing AI in healthcare, but building trust among the general public is another major challenge. According to a survey conducted by Pew Research, 60% of Americans are uncomfortable with healthcare providers relying heavily on AI. This sentiment is shared by Eric Topol, a well-known cardiologist and medfluencer, who emphasises the importance of human oversight and empathetic care in medicine. But I personally don't think humans have a monopoly on empathy. Moreover, technology can play an essential role in being empathic.

Recent research demonstrated how AI can significantly enhance care aspects previously thought to rely solely on human capabilities. Contrary to Eric Topol's assertion that "human" empathy will remain required, findings from UC San Diego Health suggest that AI can contribute as effectively to the empathetic aspects of healthcare. The study revealed that AI-generated drafts of spoken answers and written letters by doctors enhanced the quality of interactions between patients and healthcare providers. It improved nuanced human interactions and empathy in clinical settings. What's more, it did not reduce the response times of doctors but significantly alleviated cognitive burdens by providing a well-formulated starting point for patient communications.

This indicates that AI can extend beyond technical tasks and even increase trust in healthcare and towards AI itself. It will be a matter of balancing the involvement of AI and humans in healthcare to ensure that patients receive the best and most empathic care possible. AI will not replace human healthcare providers; however, those who do not adapt to incorporate AI may be disadvantaged. (you may have heard this one before).


In essence, the journey toward widespread AI integration in healthcare is as much about technological innovation as it is about cultural adaptation. This is again obvious from the news of recent weeks. I'll keep on following the news and try to bring meaning to these changes.

"Enjoyed the read? Share the joy! Spread the buzz of 'It's a Healthusiasm World' by forwarding this email to friends, family, or colleagues. Help us shine a spotlight on the power of cutting-edge tech across your favourite social media channels. Let's come together and amplify the impact!"

Stay in touch,


Other Healthusiasms

  • ⌚❤️ The FDA has recognized Apple's AFib History feature on the Apple Watch as a qualified Medical Device Development Tool, streamlining the data collection process for evaluating new digital health technologies. (The Verge )

  • 👩🏻‍⚕️ 👩🏿‍⚕️ 👩🏽‍⚕️ A study finds that patients treated by female doctors have better health outcomes and lower mortality rates compared to those treated by male doctors. ( Forbes )

  • 🍹 🍾 The mindful drinking trend, which encourages moderation and intentional consumption of alcohol or non-alcoholic alternatives, is gaining popularity as a wellness strategy. ( global wellness institute )

  • 💉💉 A recent study indicates that a universal vaccine capable of protecting diverse viral variants might be achievable by targeting common viral protein structures. ( Singularity Hub )

  • 🚚 📦 Simple HealthKit is partnering with Amazon to improve access to home diagnostics and chronic care through its telehealth services and testing kits. ( Fierce Healthcare )



bottom of page