The use of Artificial Intelligence (AI) in healthcare is gaining momentum—particularly through the advancement of Large Language Models (LLMs). These powerful language models promise greater efficiency, process automation, and a new level of quality in patient communication. In Germany, interest in practical applications is growing—provided that regulatory and data protection requirements are met.

AI as a Relief for the Healthcare Sector

The increasing burden on medical professionals—especially due to time-consuming documentation obligations—makes AI-powered solutions attractive to many institutions. Studies by Fraunhofer IKS show that LLMs in healthcare are capable of automatically structuring, classifying, and generating medical content. At the same time, technical safety, transparency, and legal compliance are identified as key success factors.

Clinical Use Cases of LLMs

A research project at the University Medical Center Freiburg compared four open-source German-language LLMs with commercial models like Claude-v2. The results demonstrate that open-source LLMs can be effectively used for medical documentation, anamnesis, or report generation—provided they are properly trained and embedded.

A survey of over 100 physicians at German university hospitals confirms this trend: many are already using AI tools for summarization, translation, or information retrieval. However, lack of integration, data privacy concerns, and regulatory uncertainty remain major barriers.

LLMs and AI in the Legal Framework

The use of LLMs in healthcare is subject to strict requirements in the EU. The most relevant frameworks include:

• the General Data Protection Regulation (GDPR),

• the Medical Device Regulation (MDR),

• and the upcoming EU AI Act.
  
  

If an LLM-based application supports medical decisions or processes health-related data, strict requirements apply regarding consent, data security, and traceability. Without these, compliant deployment in healthcare is not feasible.

ACTIMI: Secure AI Modules for Healthcare

ACTIMI integrates AI and LLMs modularly into a certified platform architecture. Our API-first approach enables the privacy-compliant use of language models within real-world care structures—with full transparency and control. Two publicly funded projects illustrate our approach:

EmpAItica

In the EmpaAItica project, we are developing an AI-supported mental health assistant for users with anxiety disorders or depression. LLMs support natural communication with a digital therapy coach.

DaDriv

As part of the DaDriv project, ACTIMI uses LLMs to generate personalised care plans for stroke patients based on medical history data. We combine rule-based AI with digital patient twins to automate individualised care paths - a concrete contribution to reducing the burden on the healthcare system.

Conclusion

LLMs and AI in healthcare are more than a trend: they offer real solutions to persistent challenges—from documentation to care planning. ACTIMI brings these technologies into practice responsibly: regulatorily compliant, modularly integrated, and publicly funded.

If you’d like to learn how ACTIMI can help you securely integrate LLMs into your digital healthcare ecosystem, feel free to get in touch.