Developing AI models in healthcare that are inherently fair, safe, and equitable for all patient populations emphasises the need for a proactive and design-first approach to defining AI problems with an equity mindset, utilising fairness-aware machine learning techniques, conducting thorough disaggregated evaluations, and establishing equitable organisational and cultural AI behaviours to prevent unintended negative bias consequences and helping to foster greater trust in AI models and systems.
Bias in healthcare AI models can lead to harmful disparities, thus it is essential to move beyond merely reacting to and mitigating bias after an AI system is developed, and instead, to intentionally design for equity from the very beginning.

In this webinar, the Dedalus team will explore:

-The types of bias and their impact in healthcare AI and the strategies to detect and reduce them.

-Equitable AI design principles with respect to fairness, accountability, transparency, and inclusive design.

-The application of pre-processing, in-processing, and post-processing methods to mitigate bias concerns in AI models.

-The importance of having diverse teams and including leadership buy-in to ensure organisational and cultural shifts for embedding an “equity by design” mindset.