Скачать или смотреть DOCSF21: Bias in AI and Algorithmic Trust

Artificial intelligence. When we hear that phrase, we tend to focus on the second word: Intelligence. We sometimes don’t spend enough time on the first word: artificial. The more data the intelligence has to work with, the more accurate it’s predictions are. But while health data is growing exponentially it is not as yet a complete or completely accurate data set. Increasingly, we are seeing tools from the world of AI being used in medicine to do everything from predicting risk of readmission to reading clinical charts for hidden diagnostic clues. But can we trust an otherwise perfectly good algorithm that is using an incomplete, partial or inaccurate data set? How big a does a data set have to be to be big enough that we can trust it? And what data sources are trust-worthy? As AI becomes ubiquitous how can we become smart consumers of it’s recommendations? How should we approach, evaluate and use AI in healthcare?

Register: https://docsf.health/annual-experienc...