Shabani observes that AI systems are increasingly taking on responsibilities traditionally held by healthcare professionals. This shift affects the relationship between patients and providers and raises questions about liability, trust and fairness. At the same time, not all groups benefit equally from these technologies. Access to digital infrastructure, levels of digital literacy and age matter, as does the way AI systems are developed. Many medical AI tools are trained on datasets that largely consist of people of European ancestry, which can make outcomes less reliable for other populations.

According to Shabani, this can result in poorer care for minority groups and erode trust in the healthcare system. When people realise that AI tools do not work equally well for everyone, they may lose confidence in healthcare altogether and even avoid seeking care. Trust, she argues, is a fundamental condition for effective healthcare, and AI should reinforce that trust rather than undermine it.

In her work, Shabani focuses on how these challenges can be addressed from a legal and regulatory perspective. She argues that regulators should play a stronger role in assessing medical AI systems. Issues of inclusivity and discrimination should be explicitly considered in approval procedures, for example through mandatory fairness checks. A clearer legal understanding of fairness is central to this approach.

The aim is to develop a practical, legally grounded framework that regulators and developers can use to make AI in healthcare more inclusive. Shabani also seeks to raise awareness and stimulate a broader societal debate on the implications of data-driven technologies in healthcare.