Life Sciences industry is extremely excited about AI capabilities and rightly so, however there’s a regulatory picture that we need to keep in mind. After all, an organization is investing time, money and resources in AI adoption with a definite ROI in mind and it is of no good if the product/service does not get regulatory approval.
AI-based systems are quite dissimilar from traditional systems in the way that it is not always possible to explain how and why the system decided in a certain vein. Most of the ML algorithms work as a Black Box, which poses an ethical issue especially, when dealing with Personal Information (PI)/Sensitive PI. This does not sit well with the regulators who need to understand how a certain conclusion was arrived. A Black Box makes an AI solution less auditable and non-transparent.
It is imperative to rationally evaluate current Risk Management practices and policies, identify the shortcomings and plug the gaps with robust technical and procedural controls.