Europe is facing an unprecedented health challenge. We are a continent with an increasingly aging population that is putting significant pressure on public healthcare systems, increasing costs and requiring innovative solutions. While AI has demonstrated extraordinary potential in areas such as diagnostics, clinical decision support, drug discovery and preventative medicine, up to 80% of AI projects in healthcare fail to move beyond the pilot phase. Or rather: it becomes difficult to move from the project to its adoption.
The point on the topic was made during the conference “AI for Healthcare – Longevity & Wellness”an international conference held in Genoa organized by Fusion AI Labs, a center of excellence that promotes applied research and industrial experimentation on artificial intelligence. The meeting is part of the European strategies “Apply AI” and “AI in Science”, promoted by the European Commission to accelerate the transition from research to concrete implementation: the first focused on adoption in healthcare systems, the second on strengthening scientific infrastructures.
The future that awaits us
Europe has an extraordinary opportunity and also a responsibility. The healthcare sector is under structural pressure: in Italy we have 64 nurses per 10,000 inhabitants, compared to a European average of 80, and just 1.5 nurses per doctor compared to 2.2 in the rest of Europe. 30% of European doctors are over 55 years old.
“It is not a crisis that can be solved just by hiring more staff, but a crisis that requires a paradigm shift in how we organize, support and empower the entire system”
comments Andrea Pescino, CEO of Fusion AI Labs, outlining the crucial transition from innovation to concrete impact.
“Applied research on artificial intelligence in the healthcare sector cannot be limited to clinical applications, however extraordinary. Let’s think about the value of accelerating operational processes, optimizing the management of departments, improving the quality of interaction between facilities and patients. Every hour saved in an administrative process is an hour returned to care. Every system that better supports a nurse or doctor in managing the daily load reduces burnout and increases the quality of care”.
In this sense, the event brought together a broad ecosystem of public and private actors: European and national institutions, international organisations, research centres, universities, technology companies, healthcare and pharmaceutical sectors. Participants included representatives of the European Parliament, organizations such as UNESCO and the World Bank, as well as large global companies and centers of excellence such as Imperial College, Karolinska and Charité.
AI strategies and applications
We must not think of AI as a sort of “medical superintelligence” capable of promoting simple diagnostic improvement. Antonio Gatti – Global Lead Pharma Life Sciences Microsoft is convinced of this:
“Frontier models support longitudinal reasoning, integrate multimodal data – omics, imaging, medical records, real-world evidence – orchestrate tools, maintain memory. For the first time it is possible to think of AI as an infrastructure of the healthcare and pharmaceutical pipeline, not as a single application. Yet up to 80% of AI projects in healthcare do not pass the pilot phase: not due to limitations of the technology, but due to how it meets the system – data, flows, regulation, trust. The leap in quality is not on the model, it is on the context. From the discovery of new therapeutic targets to the design of clinical trials, from patient stratification to precision medicine: in a Europe with fewer doctors, fewer nurses and older people than a decade ago, medical superintelligence is not a technological option but a sustainability lever for the entire life sciences supply chain. The question is not whether to adopt it, but how: responsibly, on scale, oriented towards real impact”.
The proof comes from what was reported by Kiril Veselkov, associate professor in Computational Medicine and Oncology Informatics at Imperial College and Antonello Scalmato, CTO of Fusion AI Labs. The experts presented the advanced applications of generative AI in the early diagnosis of gastric cancer conducted with the European research project AIDA. The intervention presents an integrated vision for precision gastric cancer prevention, combining AI-based diagnostics, personalized treatments and nutrition-based prevention strategies.
“It highlights work on AI applied to the assessment of gastric inflammation, including the H. pylori AI Clinician to support personalized treatment decisions, as well as data-driven approaches to understanding how dietary factors, metabolism and biomarkers interact within biological networks linked to disease progression. Taken together, these approaches show how AI can connect diagnosis, therapy and personalized prevention, helping to reduce the risk of progression to gastric cancer.”
is Veselkov’s comment.
AI for early diagnosis
Paula Petrone, head of the Digital Health Unit at the Barcelona Supercomputing Center (BSC), is leading the development of AI-based solutions for early disease diagnosis, risk assessment and personalized treatments, using multimodal health data, including medical images, wearable devices and electronic health records.
In his speech he took stock of the paradoxes of health and well-being in Europe and the opportunities offered by AI. The data speaks clearly: we live longer, but not necessarily better. Over the past twenty years, life expectancy has increased significantly, while years lived in good health have remained stable, widening the gap between longevity and well-being. At the root of this trend are insufficient investments in prevention and a fragmented healthcare system, little oriented towards the integrated management of chronic diseases. Be careful though.
“However, the study highlights the potential of generative artificial intelligence and initiatives such as the European Health Data Area in promoting more preventative and coordinated care models. It remains essential to address the technological, ethical and governance challenges to ensure equitable and sustainable access to the benefits of these innovations”
summarized the expert.
Support for robotics
Fulvio Mastrogiovanni, associate professor at the University of Genoa, explored the integration between AI and robotics in the healthcare sector, exploring the ongoing transition from purely digital artificial intelligence to systems capable of physically interacting with the real world: Embodied AI.
“The first part clarifies what “embodiment” really means. Not simply models that process data, but agents that perceive, act and learn through a body”
explains Mastrogiovanni.
Recalling recent research published on Science on the role of touch and advanced sensing systems, the speech highlights how AI capabilities must be rethought when moving from digital environments to the physical world, where uncertainty, security and human interaction become central. For the rest, the expert proposed a structured overview of applications in the real world along an evolutionary spectrum: from pick-and-place systems and robotic manipulation to assistive robotics, up to advanced areas such as robotic surgery, intelligent prosthetics and companion systems.
