How can AI help reduce pressure on our healthcare system?

Current applications of AI in healthcare

Traditional AI is already widely used in Dutch healthcare institutions, particularly in radiology for image recognition and in top clinical hospitals for diagnosis, treatment, and prognosis. For example, Amsterdam UMC is conducting studies to better predict cardiovascular diseases using AI. Generative AI—capable of learning from input data and creating new data—is still in its early stages and mainly used in research projects. The expectation is that GAI can help reduce the workload of healthcare professionals and support physicians and pharmaceutical developers.

Opportunities to support healthcare professionals

One of the most immediate opportunities is reducing the administrative burden on healthcare providers. GAI can assist with answering patient questions and summarizing medical records, freeing up time for direct patient care. Another key application is supporting physicians in diagnosis and treatment planning, for example by quickly extracting relevant information from medical files. This can improve both efficiency and accuracy in clinical decision-making.

Opportunities for personalized care

In the longer term, GAI can enable more personalized care through an integrated approach. It can rapidly collect, analyze, and enrich all available medical data, helping to create care plans tailored to each patient’s unique needs. This is particularly relevant for elderly care, where patients often require multiple types of treatment and costs are high.

Accelerating drug development

GAI can also play a major role in drug discovery and development, improving treatment options. For example, it can identify new properties of existing molecules, generate new molecular structures, and predict clinical outcomes—allowing the pharmaceutical industry to develop medicines faster.

Risks: errors and dependence on big tech

While GAI offers great potential, TNO researchers highlight significant challenges. A major risk is “hallucinations”, where AI generates incorrect or misleading information, which could have serious consequences. Another concern is dependence on large technology companies that develop GAI technologies, potentially weakening the negotiating position of healthcare organizations and increasing costs.

Challenges: legislation and privacy

Regulations governing the use of GAI in healthcare are often unclear or lag behind technological developments, creating legal uncertainty and slowing implementation. Ethical and privacy considerations must also be addressed—especially the confidentiality of patient data, which is essential for maintaining trust among patients and healthcare providers.

Recommendations for implementing generative AI

A key finding from TNO’s analysis is that human oversight is critical to ensure the safety, integrity, and reliability of GAI in healthcare. Additional measures include reducing dependency on foreign tech companies and aligning innovations with Dutch healthcare practices—for example, by investing in domestic GAI development and promoting collaboration between healthcare, technology, and government. Training healthcare professionals is also essential to familiarize them with GAI’s capabilities, limitations, and ethical implications.

Accelerating AI implementation together

By taking the right steps, GAI can become a valuable addition to healthcare, contributing to better, more efficient, and more personalized care. TNO can play a role in reducing fragmentation and accelerating the adoption of GAI in healthcare.