Medical Research: "Behind Every Data Point, There's a Human"
Dr. Jacqueline Lammert has outlined key challenges and solutions for integrating AI into healthcare. Her insights focus on improving data quality, infrastructure, and human oversight to build trust in medical AI systems. She also highlighted major industry players and the importance of European digital sovereignty in the sector.
Dr. Lammert pointed to poor data quality as a critical obstacle, noting that over 80% of healthcare data remains unstructured. Despite this, her team has successfully used large language models (LLMs) to extract diagnoses, therapies, and biomarker profiles from text-based records.
She stressed that trust in AI depends on involving humans from the earliest stages of development. While acknowledging the potential of LLMs, she urged caution, calling for better personnel training and stronger risk awareness in their deployment. Infrastructure expansion was another priority. Dr. Lammert argued for a secure European cloud network capable of real-time data processing in medicine. She also advocated for open-source standards like Kubernetes and core datasets to ensure smooth data exchange between systems. Among the leading providers in AI hardware and cloud services, she named Microsoft, Google, Amazon (AWS), IBM, and Oracle. Additionally, Dr. Lammert supports open standards and open-source software as essential for maintaining Europe’s digital independence in healthcare. Her work includes the GoTwin project, which uses digital twins to create personalised therapies for ovarian cancer patients.
Dr. Lammert’s recommendations cover data standardisation, infrastructure upgrades, and human-centred AI development. The GoTwin project demonstrates how digital twins could transform cancer treatment. Her approach combines technological innovation with strict safeguards to ensure reliability in medical AI applications.
