There has been a growing interest in artificial intelligence (AI) during recent years, and multiple studies have assessed the real-world clinical experience of radiologists with AI [1], [2], [3], [4], [5], [6]. While many AI-based solutions are commercially available, it is challenging for radiologists to assess their qualities and understand the expectations related to these new techniques, especially in the context of explaining results to patients [7]. Radiologists need to continue making medical decisions, regardless of the accuracy of AI predictions, to maintain patient trust [8,9]. Therefore, it is crucial for radiologists to be aware of the limits of AI software to prevent overuse and misuse of AI in a background which could lead to patient harm.

Radiological initiatives worldwide have established AI working groups to discuss and deliberate on the introduction and implementation of AI in imaging, aiming to inform radiologists about AI solutions [10]. DRIM France AI, created in 2018 as a non-profit association, supports the development of AI in radiology, helping radiologists to choose relevant and validated AI solutions, ensuring compliance with regulations, and meeting the needs of French radiologists [11].

The objective of this study, initiated by the radiology community, was to create and validate the first French descriptive and analytical grid of radiological AI solutions.