The Implementing AI in healthcare is slow compared to other sectors. To corona time. "In one week we developed a planning tool. It just had to be done." This zegt Jeroen de Haas, founder and CEO of Pipple, during the masterclass 'Data and AI for healthcare', organized by Dutch Health Hub and Health~Holland.
What's on your mind?
As an econometrician, guitarist and DJ, Jeroen de Haas is at home in many fields. Ten years ago, he founded Pipple, a data agency with a team of mathematicians and engineers that specializes in solving complex problems through data and AI. De Haas describes Pipple as creative and open-minded, with a focus on people, planet and profit.
What's on your mind? This is how De Haas begins his master class. The flipchart fills with hopes and expectations from the audience. How do you handle data exchange? How do you ensure data quality? How do you avoid distrust of AI? How do you go from pilot to long-term implementation? And: what is the future perspective?
Ownership of data
Without dates, you're just another person with an opinion, a quote from W. Edwards Deming, appeart on screen. A conversation about data ensues and possession. Data are not from someone, but they are going about someone. De Haas nuances: "Data is increasingly becoming an asset. And every asset needs an owner." Ownership and responsibility, he says, are crucial for reliable deployment of AI inside organizations.
The failure of AI projects
AI is plentiful in the public debate, but what will remain when the hype fades? What has real potential and what is hot air? De Haas sees that right now, organizations are often not ready for successful implementation. According to him, AI projects do not await a fruitful future for the following five reasons:
-
t
- De verkeerde mensen op cruciale (kostbare) posities
- Slechte datakwaliteit
- Het onderschatten van de werkelijkheid
- IT als bottleneck met een te lange time-to-market
- Weerstand tegen verandering
t
t
t
t
Data quality in particular is a big problem, according to De Haas. "People generally don't like entering data. I don't know many people who do it as a hobby." Embracing change also takes a lot of energy. Why are we doing it? What do we gain from it? "That's often approached too technically.
Inertia in healthcare
Implementing AI in healthcare is slow compared to other sectors, cited. The Haas. But sometimes things can suddenly move fast. "In 2020 We did a strategy project for the Jeroen Bosch Hospital. The report disappeared into the drawer. Until corona broke out. The influx of patients was completely unpredictable. In one week, we developed a scheduling tool. It just had to. That shows: if it has to, it can be done."
Ethics and loss of professionalism
AI is more than technology: it also gets to the heart of healthcare professionalism. The Haas tellt about his neighbor those working is in mental health. She uses as a trial a chatbot which summarizes patient records to start the conversation from there. Useful, but also problematic. What information is being filtered? And in doing so, aren't you giving away a piece of your craftsmanship? What do we actually want AI to do in healthcare? And what would rather not?
Trust and adoption
To go distrust of AI in healthcare, De Haas stresses the importance of co-creation: "Include healthcare professionals in the development. Analysts don't know reality, the shop floor does." Only then can a pilot grow into sustainable adoption.
A practical example of continued use comes from home care organization Tzorg, where Pipple developed a scheduling solution. "Legislation, technology and business processes are constantly changing. You're never ready. Building and transferring something externally does not work if there is no capacity internally to continue developing." Therefore, Pipple also strives to be the number one partner in data and AI, especially in the long term.
Politics and power
Politics and governments can also play a big role. Decisions from above can inhibit innovation, but sometimes accelerate it. The Haas wijst to the financial sector, where banks are required by legislation to share data. "I can imagine that in other sectors obligations come to encourage innovation."
Future Prospect
According to The Hare It is interesting to look at how we are slowly moving from AI tools to AI-agents those perform tasks autonomously and make choices. When it goes talking about the future then also mostly about Artificial General Intelligence (AGI) and Superintelligence. "But if and when that will happen, I can't say anything about that right now."
This masterclass laat see that AI in healthcare is not about technology alone. It's about ownership, ethics, trust and change capacity. Or like The Hare summary: “Adaptability and risk appetite of organizations play an important role when it comes to implementing AI.”
The next masterclass 'Data and AI for healthcare' is on Wednesday, October 1. at Jaarbeurs Utrecht. Curious about the experiences and stories of Wouter Kroese of Pacmed and Patrick Leenen of Agentic Alliance? Then sign up by sending an email to Ruud Koolen at ruud.koolen@jaarbeurs.nl, stating your name, position, organization and phone number.
