During TeamUp X Health, an event that gives healthtech startups a platform to pitch their ideas and connect with potential customers, employees and investors, a panel discussion named “AI in Healthcare” took place.
Smion organized the event within the AI2MED project which provided insights into the current applications of artificial intelligence (AI) in medicine, highlighting how many technologies have transitioned from theory into daily practice.
This meetup provided a unique opportunity to connect with experts in the healthcare sector and find inspiration and opportunities for future collaborations in applying AI in the healthcare industry, and the goals align with those of AI2MED. Namely, the project aims to integrate AI into medical practice, optimize diagnostics and treatment, and bridge the skills gap through training and collaboration. It seeks to transform healthcare, making it more precise, efficient, and accessible.
The discussion featured experts in the healthcare field: Ino Protrka, MD, Director of DZ Zagreb Centar; Klara Zubčić, PhD, a molecular biologist from Dubrava Clinical Hospital; Prof. Leo Mršić, PhD, Vice-Rector for Science and Research at Algebra University; and Željko Švedić, founder of TestDome and GemBox. The panel was moderated by Mladen Kovaček, founder of the Future Is Now conference.
AI in Diagnostics
Klara Zubčić, PhD emphasized that artificial intelligence already plays a pivotal role in molecular biology, particularly in gene sequencing and genome analysis. Thanks to AI tools, processes that once took weeks can now be completed in just a few days. “AI is already in use; we have surpassed the potential,” Zubčić noted, adding that she uses software based on AI models in her daily diagnostic and research work.
One of the most recent examples of AI’s impact is the awarding of the Nobel Prize in Chemistry for DeepMind’s AlphaFold model, which predicts protein structures. Such advanced AI tools accelerate research and pave the way for the development of new therapies.
Legislative Challenges and Limitations
Ino Protrka, MD addressed the legislative barriers that limit AI applications in healthcare, despite Croatia’s robust technological infrastructure, such as the CEZIH system. Comparing Croatia to other European countries, he pointed out that Croatian systems have a significantly larger amount of data but are constrained by regulatory frameworks. “AI is still underutilized in primary healthcare, even though we have some of the best conditions in Europe,” said Protrka.
Reliability and Transparency of AI Models
Željko Švedić identified “truthfulness” and “explainability” as two critical issues in the application of artificial intelligence in healthcare. Questions arise about how reliable the results of AI models are and whether physicians can understand the reasoning behind these conclusions. He emphasized that diagnoses suggested by AI models must always undergo expert verification. Additionally, the panel raised the question of accountability in cases of incorrect AI assessments.
Role of Academia in Triple Helix Ecosystem (Alongside Industry and Government)
Prof. Leo Mršić, PhD highlighted the importance of collaboration between academia, industry, and the government in developing AI solutions for healthcare. Academia needs to act as a bridge, catalyzing synergies in pursuit for effective and reliable solutions. Mršić pointed out importance and opportunity for Croatia to support local ecosystem engaged in developing innovations in healthcare.
However, the panel also pointed out a gap in the education system, which has yet to keep pace with the rapid advancements in AI technology. While programming courses are available, medical faculties are slow in providing the informatics education needed for working with AI tools. One proposal was to introduce a course titled “Artificial Medicine,” designed to ensure that both students and professors continuously update their knowledge to keep up with the fast-evolving technology.
Data as a Key Resource
Panelists emphasized that high-quality data is critical for developing AI-driven diagnostics. However, there are questions about whether Croatia has a sufficiently large and high-quality data pool to enable the broad application of these technologies. Improved access to and exchange of data between institutions could significantly enhance diagnostic accuracy and decision-making speed.
Conclusion
The discussion demonstrated that artificial intelligence is already an integral part of modern healthcare. However, fully realizing its potential requires addressing legislative, educational, and infrastructural barriers. Given its existing resources and infrastructure, Croatia has the opportunity to take a leading role in AI adoption in healthcare. Achieving this, however, will require stronger synergy among all key stakeholders, including physicians, scientists, policymakers, and technology experts.
