In How Many Ways Can an Algorithm be Fair?

Explore the complex landscape of algorithmic fairness in this 38-minute talk by Suchana Seth at the Alan Turing Institute. Delve into various measures of fairness in predictive algorithms and their implications for technology policy and regulation. Examine the challenges in implementing these fairness measures and learn how they can be used to hold algorithms accountable. Gain insights from Seth's expertise as a physicist-turned-data scientist, covering topics such as algorithmic bias, stakeholder consensus, accuracy tradeoffs, and real-world examples like Airbnb. Discover the intersection of ethics and machine learning, and understand the importance of fairness, accountability, and transparency in AI systems. Engage with the evolving regulatory landscape for predictive algorithms and consider the broader implications for the future of ethical AI development and implementation.

Intro
What is algorithmic bias
Challenges
Stakeholder Consensus
Mutually Exclusive Measures
Accuracy Tradeoffs
Algorithms
Airbnb
Accountability mechanisms
Policy implications
Questions