AI in health sector grows

Artificial intelligence is becoming more common in healthcare, especially in analyzing data, helping with diagnoses, and interpreting medical images. However, its use among nurses is still limited, and discussions about its role are in early stages.

Nursing and AI challenges

Flavie Aroulandom, a senior health officer and pediatric nurse at the Assistance Publique - Hôpitaux de Paris (AP-HP), says there is no real change in nursing due to AI yet. She points to high costs, the need for evaluation, compatibility with hospital systems, and compliance with cybersecurity rules as major barriers.

Efficiency and shadow AI

AI is seen as a way to make nursing work more efficient by reducing time on routine tasks and making information easier to find. Some nurses use AI platforms without official approval, a practice called "shadow AI." Flavie Aroulandom acknowledges this but warns of risks like misuse and possible leaks of patient data.

Private sector AI use

In the private sector, nurses also use AI, though they don't usually call it "shadow AI" because they work independently. Julien Grizzetti, a private nurse and consultant for software publishers, says AI helps in researching drug effects, understanding patient conditions, and interpreting imaging reports more quickly than traditional methods.

Need for critical use of AI

Julien Grizzetti warns that AI is not yet reliable in areas like prescription valuation, where errors and non-standardized prescriptions can lead to wrong results. He emphasizes the importance of critically analyzing AI outputs and not relying solely on its responses.

Training and awareness

Flavie Aroulandom stresses the need for training and awareness about AI, including understanding risks, cybersecurity, and the protection of health data. She says education is crucial to help professionals detect inconsistencies and question the sources in AI-generated answers.

Custom AI tools for nurses

Both Flavie Aroulandom and Julien Grizzetti agree that AI tools should be tailored to the needs of nurses. Flavie Aroulandom has developed a prototype tool that searches hospital protocols and guidelines, ensuring it doesn’t make up answers. The tool needs constant human updates and is currently in the proof-of-concept phase.

Private sector AI integration

On the private side, software publishers are adding AI to their services, such as scheduling appointments and managing accounting, using closed systems that don’t share patient data. Julien Grizzetti notes that automatic valuation is still difficult because of the complexity of integrating patient-specific data into AI tools. He suggests that manual methods may still be more practical in some cases.