Install SubjectCards App

Add to your home screen for better experience

Data Science & AI

Weapons of Math Destruction

by Cathy O'Neil

5
Key Concepts
5
Action Items
1
Core Thesis
1
Mindset Shift

Key Concepts

1

WDM Criteria

Models are destructive when opaque, unregulated, and unfair, causing widespread damage.

2

Algorithmic Bias

Algorithms reflect and amplify human biases present in their training data, leading to discriminatory outcomes.

3

Feedback Loops

Models can create self-fulfilling prophecies, trapping individuals in negative cycles based on their predictions.

4

Proxy Variables

Seemingly neutral data points often stand in for sensitive, discriminatory characteristics, perpetuating bias.

5

Opacity Problem

The lack of transparency in complex algorithmic decisions prevents accountability and challenge, hiding their impact.

Weapons of Math Destruction by Cathy O'Neil
Get Book

subjectcards.com/books

Action Items

Demand transparency and explainability in all algorithmic systems.

Rigorously audit models for bias and fairness before and after deployment.

Implement robust human oversight and appeal mechanisms for automated decisions.

Actively challenge the use of discriminatory proxy variables in model design.

Advocate for strong ethical guidelines and regulatory frameworks for AI.

Core Thesis

Unchecked algorithms, fueled by big data, can become Weapons of Math Destruction that perpetuate and amplify societal inequality.

Mindset Shift

Data science is not inherently objective; it is a powerful tool that can amplify existing injustices if not ethically designed and critically scrutinized.

SubjectCards.com Discover • Learn • Grow