AI to address discrimination

Using machine learning to highlight in real term where a judgement and/or decision is a risk of or may have been subject to bias.  There is a sustained focus upon the risk of machine learning might exhibit the bias of its creator but less attention has been paid to the extent to which AI can signal where patterns of bias exisit within health-related decision making. There is evidence that indicates that healthcare practitioners  possess identical levels of bias as the wider population but that such bias is not necessarily conscious or subject to scrutiny.  Utilising healthcare data sets the proposal is to use AI to signal where a particular decision or set of decisions are at risk of bias.  This would enable the healthcare profession to address the role of implicit biases within healtcare decisions in real-time. 

Why the contribution is important

Discrimination means that a person or group is treated less favourably than another person or group based on their background and/or certain personal characteristics resulting in a loss of access to goods or services.  The focus here is upon healthcare discrimination. Individuals should not expect to receive a sub-optimal care because of their race, age, disability or any other characteristic. Perceptions of discrimination act as a cogntive stressor which research shows reduces physical and mental well-being over time.  Using technology to signal where bias may existing will help physicians and health providers to review and interrogate decisions on an ongoing basis and increase awareness and sensitivity to discrimination and ensure that the public approach healthcare with confidence in being treated without prejudice.  

by DrWaite on July 20, 2018 at 04:53PM

Current Rating

Average rating: 0.0
Based on: 0 votes