Dr Olalekan Ajayi: AI Models Are Only as Good as Their Data

In this interview from the Association of Community Cancer Centers (ACCC) 2023 Annual Meeting and Cancer Center Business Summit, Olalekan Ajayi, PharmD, MBA, ACCC president and chief operating officer of Highlands Oncology Group, PA, in Fayetteville, Arkansas, addresses the use of artificial intelligence (AI) in health care equity and why data integrity is just as important as the models used with the data, in order to best optimize patient interventions.

Transcript

What are some best practices for optimizing management of AI that can start to break down health care equity gaps?

So, AI is still in its infancy as far as health care goes. One thing we're learning about AI is that the models are only as good as the data that you train them on. And so, one of the far more important things would be the integrity of the data that we're using to train these models and what controls can we establish to ensure that these models are trained in the right way.

The other thing I would also say is that the models have the ability to show us where to focus on—particularly in equity—so we can understand if people are at risk, and we can make the interventions ahead of time. We have the ability to predict, which is really exciting. But again, it will go back to a very clear understanding of what the model is saying, just so that we're having the right interventions for the right populations.