Webinar on IT. Transparency and Interoperability in AI developments in healthcare

The widespread integration of Artificial Intelligence (AI) algorithms and models into the forefront of clinical practice remains a challenging endeavour, primarily dominated by major corporations. This predicament stems from several significant hurdles, with interoperability and transparency as pivotal obstacles.

Ensuring interoperability is crucial across all stages of implementing AI in healthcare. Developing accurate AI models requires extensive data, often unavailable within a single healthcare setting. Consequently, it becomes imperative to amalgamate datasets from various sources or employ federated learning approaches. Even when a dataset is accessible during the training phase, it is essential to cross-validate the model on diverse cohorts to assess its generalizability and safety. This importance heightens upon the model’s readiness, necessitating a flexible business model and a time- and cost-efficient method for seamlessly integrating the solution into diverse healthcare settings for practical real-world use. Transparency in AI is even more crucial in healthcare as it is directly related to the reliability and safety of patients in addition to being a precondition for explainable and reproducible AI research.

Within the AICCELERATE project, we offer a methodology and a suite of tools that leverage health data standards such as HL7 FHIR to address challenges from a data perspective. In this webinar, we will provide a concise overview of this approach and discuss how technology and standardization can serve as effective solutions to tackle these issues.

Did you miss it?