Presentation Type
Invited
Student
No
Track
Methodology
Abstract
Alarming headlines appear at an increasing rate about ‘biased AI’ leading to nontransparent decisions, discrimination and lack of trust.
Newcomers to AI application development are confronted with (and often struggle with) multiple traps that can inadvertently lead to biased and unreasonable automated decisions. By comparison, credit score developers have decades of experience in the “art and science” of mitigating biases through transparent and credible models and decisions that provide benefits to all stakeholders while complying with fair lending regulations.
In this presentation you will learn about traps to be avoided and get our unique perspective on how to achieve comprehension and trust while beating back biases through a special synthesis of data, expertise and technology. The battle to minimize biases is ongoing and we will also share our latest thinking how we can up the ante in this endeavor.
Start Date
2-11-2020 3:30 PM
End Date
2-11-2020 4:30 PM
Included in
Business Analytics Commons, Business Law, Public Responsibility, and Ethics Commons, Finance and Financial Management Commons
Session 12 - Finance: Bias In, Bias Out? How to Fight Biases in AI and Create Trust
Pheasant Room 253 A/B
Alarming headlines appear at an increasing rate about ‘biased AI’ leading to nontransparent decisions, discrimination and lack of trust.
Newcomers to AI application development are confronted with (and often struggle with) multiple traps that can inadvertently lead to biased and unreasonable automated decisions. By comparison, credit score developers have decades of experience in the “art and science” of mitigating biases through transparent and credible models and decisions that provide benefits to all stakeholders while complying with fair lending regulations.
In this presentation you will learn about traps to be avoided and get our unique perspective on how to achieve comprehension and trust while beating back biases through a special synthesis of data, expertise and technology. The battle to minimize biases is ongoing and we will also share our latest thinking how we can up the ante in this endeavor.