Chapter
The Complex Definition of 'Fairness' in the Tech Industry
27:49 - 36:43 (08:54)
The tech industry has developed a notion of fairness that takes into account ethical and legal fairness, which includes whether different groups of people are affected differently by a machine learning system. However, there are alternative definitions that consider if the model makes the same kinds of errors for different groups of people and how civil rights legislation applies to statistical analysis.
Clips
The field of computer science has become preoccupied with fairness, not only in terms of equal access to technology but also taking into account different groups of people being affected differently by machine learning systems.
27:49 - 29:39 (01:50)
Summary
The field of computer science has become preoccupied with fairness, not only in terms of equal access to technology but also taking into account different groups of people being affected differently by machine learning systems. This is especially seen when it comes to decisions such as determining whether an individual should be held in jail before their trial or released to go home before their trial.
ChapterThe Complex Definition of 'Fairness' in the Tech Industry
Episode#297 - Brian Christian - The Alignment Problem: AI's Scary Challenge
PodcastModern Wisdom
Cynthia Rudin explains the impact of implicit bias in AI and machine learning, highlighting the issues of accuracy and fairness that can arise from a lack of diversity in developers and data used for algorithms.
29:39 - 34:31 (04:52)
Summary
Cynthia Rudin explains the impact of implicit bias in AI and machine learning, highlighting the issues of accuracy and fairness that can arise from a lack of diversity in developers and data used for algorithms. She argues for a public policy conversation to decide the parameters of when AI is being used most effectively and fairly.
ChapterThe Complex Definition of 'Fairness' in the Tech Industry
Episode#297 - Brian Christian - The Alignment Problem: AI's Scary Challenge
PodcastModern Wisdom
Predictive policing attempts to measure crime but often ends up perpetuating bias in the system.
34:31 - 36:43 (02:12)
Summary
Predictive policing attempts to measure crime but often ends up perpetuating bias in the system. The flaws with this technology can lead to disparities in arrest based on race rather than actual crime data.