Page 19 - Shaping ethics, regulation and standardization in AI for health

Shaping ethics, regulation and standardization in AI for health

Figure 3: Example of OCI building block: Evaluation

– Evaluation (audit) package: Offers testing measures and methods for different quality
dimensions including interpretation, bias, uncertainty, and robustness; questionnaires
provide qualitative evaluation addressing ethical and governance considerations.
– Reporting package: A customizable reporting interface that presents the results of the
Evaluation package.

Over 40 developers, regulators, and medical professionals from five continents were involved
in the OCI development during the FG-AI4H lifetime.

To timely deliver resilient software capability the OCI is implementing strategies that focus
on cybersecurity and survivability across the development process. The OCI adopted the
Development, Security, and Operations (DevSecOps) approach, which is a widely-adopted
industry framework that integrates security into every phase of the software development
lifecycle and enables the delivery of resilient software in a short development cycle. The
DevSecOps software lifecycle approach creates cross-functional teams that unify historically
disparate evolutions: development, cybersecurity, and operations. As a unified team, they
follow agile design principles and embrace a culture that recognizes resilient software is only
possible at the intersection of quality, stability, and security.

The platform prototype was tested in different proof-of concept activities and a specification
has been developed for data annotation in health in FG-AI4H DEL 5.3 [8]. The integration
process of Topic Group TG-Symptoms, TG-Fall, and TG-Neuro with the OCI platform yielded
significant enhancements to the Data Acquisition and Data Annotation packages. Notably,
their inputs were instrumental in refining these packages. The validation of the assessment
platform and its associated processes was conducted with active participation from several TGs

9