AI and Robotics in Healthcare: Insights from dScience’s Session at Norway Life Science Conference 2025
Oslo, February 12, 2025 – At the Norway Life Science Conference 2025, the parallel session “Technology Outlook – The Future of AI and Robotics for the Health Sector” brought together experts, researchers, and healthcare professionals at Meet Ullev?l. The session examined the growing impact of AI and robotics in healthcare, highlighting key challenges and future directions.
The session was chaired by Professor Geir Kjetil Sandve from the University of Oslo.
Chaired by Professor Geir Kjetil Sandve from the University of Oslo, the session covered a range of topics, including AI-driven personalised treatment and the role of robotics in surgeries. Speakers highlighted the potential benefits of these technologies while addressing the practical challenges that need to be resolved for broader implementation.
Key Presentations and Insights
When Will a Robot Treat Us Instead of a Medical Doctor?
Jim T?rresen.
Jim T?rresen, professor specialising in AI and robotics at the University of Oslo, opened the session by exploring the evolution of robotic assistance in healthcare. He outlined how robots have already made strides in areas such as logistics, disinfection, and patient transport.
T?rresen also highlighted the research conducted by the Robotics and Intelligent Systems (ROBIN) group at the Department of Informatics, UiO.
However, the transition from support roles to direct patient interaction remains gradual, mainly due to concerns about safety, precision, and cost-effectiveness.
"It's a gradual transition rather than a binary on-off thing," T?rresen stated. "At the same time, if they are much slower than a human, we will be too restless and ask for the human instead."
Can Large Language Models Predict the Efficacy of Treatments for Individual Patients?
Kristian Svendsen, pharmacoepidemiologist from Nordic RWE.
Simen Eide, AI developer and researcher at Schibsted, and Kristian Svendsen, a pharmacoepidemiologist from Nordic RWE, presented their pioneering work on using large language models (LLMs) to predict patient outcomes. Their research involved training AI on extensive health data to forecast the likelihood of dementia development.
Simen Eide, AI developer and researcher at Schibsted.
"Registered data is just text," Eide explained. "You get born, you go to school, you get some medications at this time, and so on. This can be described as text data. So basically, that’s what we want to do—structure registry data in a way that allows AI to make predictions."
The slide from Svendsen & Eide presents findings on whether large language models (LLMs) can predict dementia:
1. Yes – LLMs can distinguish between dementia and non-dementia cases.
2. Yes – LLMs outperform statistical baseline methods such as logistic regression.
3. Todo – Further research is needed to determine:
? Whether LLMs can predict dementia over longer time periods.
? If the probability estimates produced by LLMs are properly calibrated.
They showcased real-world applications, such as AI-assisted radiology and sepsis prediction models, demonstrating how data-driven solutions can enhance early detection and improve patient outcomes.
"Many of us think that we might just be scratching the surface of what is possible," Alvaro K?hn Luque said.
AI and machine learning continue to offer unprecedented opportunities to enhance diagnostics, personalise treatments, and accelerate drug discovery. However, the road to fully integrating AI into clinical practice is not without obstacles. Addressing these challenges head-on, Valeria Vitelli emphasised:
"Machine learning is improving efficiency and accuracy of clinical trials, but we still face challenges—bias, reproducibility, data hunger, and trust."
"The future is bright, obviously, but the potential benefits also need a reality check. It needs a contact with humans. It needs a perception of knowledge," she said.
Slide text:
AI-ML in healthcare tremendously promising: bright future!
Potential benefits need a reality check
Domain human knowledge crucial for societal benefit:
It can expand our capabilities BUT not substitute humans.
AI Engineering and Integration into Healthcare IT Systems
Antonio Martini.
Antonio Martini, professor at the University of Oslo, shifted the focus to the practical side of AI deployment. He discussed the emerging field of AI engineering, emphasising the need for sustainable integration of AI into complex hospital IT infrastructures.
"Developing AI models is just the beginning," Martini remarked. "The real challenge is ensuring they remain functional and adaptable in dynamic healthcare environments."
"We need to invest in AI engineering. It’s not just about what we build, but how we roll it out and maintain it over time," he said.
Quantum Computing’s Role in Life Sciences
Lars Nordbryhn.
Lars Nordbryhn, IBM Quantum Ambassador, concluded the session with a glimpse into the future of quantum computing in healthcare. He discussed how quantum computing could revolutionise molecular modeling, protein structure prediction, and treatment personalisation.
Development Roadmap from IBM Quantum.
"Quantum computing is noisy and still developing, but we are already seeing real use cases," Nordbryhn explained.
Quantum computing impacts.
"We are opening up for collaboration and research on how to utilise quantum computing in real-world scenarios."
KEY TAKEAWAY FROM THE SESSION
A key takeaway from the session was the strong consensus on the necessity of interdisciplinary collaboration. Speakers highlighted that AI and robotics experts must work closely with medical professionals to ensure that technological advancements align with clinical needs and regulatory frameworks.
Geir Kjetil Sandve.
“The challenge is not just in developing AI models but in making them seamlessly integrate into real-world healthcare settings,” noted Professor Sandve.
“If we want real change, we need collaboration between engineers, doctors, and policymakers.”
A big thank you to all the presenters!
The session attracted significant interest, thanks to the engaging presentations. The dScience team extends our gratitude to everyone who contributed. Special thanks to UiO:Life Science and partners for organising a great conference!
