logo

67 pages 2 hours read

Brian Christian

The Alignment Problem: Machine Learning and Human Values

Brian ChristianNonfiction | Book | Adult | Published in 2020

A modern alternative to SparkNotes and CliffsNotes, SuperSummary offers high-quality Study Guides with detailed chapter summaries and analysis of major themes, characters, and more.

ConclusionChapter Summaries & Analyses

Conclusion Summary

Christian opens the Conclusion of the book with an account of his Christmas Eve spent with his wife at his father’s house. Christian awakens sweating heavily due to overheating in his room caused by a misaligned thermostat in another room, open to the colder air of the house. This incident illustrates the practical risks of relying on systems without fully understanding how they function and their potential hazards.

Christian then follows with a chapter-by-chapter overview of the book. Chapter 1 of the book discusses the importance of representative training data in model development, with a focus on the shift toward more inclusive data sets in face recognition technology. Despite progress in consumer technology, studies reveal bias against racial minorities already subject to high surveillance. The chapter also compares these data representation issues with longstanding biases in medical trials predominantly conducted on men, exposing systemic issues in both fields. These discussions, emphasize the broader challenges to ensuring that AI systems and medical practices are inclusive and fair.

Chapter 2 discusses the use of risk assessment tools in criminal justice, pointing to the fact that they often rely on problematic proxies like rearrest and reconviction rates instead of actual recidivism. These tools may reinforce biases, especially if certain demographics are more likely to be arrested or convicted.

blurred text
blurred text
blurred text
blurred text
Unlock IconUnlock all 67 pages of this Study Guide

Plus, gain access to 8,650+ more expert-written Study Guides.

Including features:

+ Mobile App
+ Printable PDF
+ Literary AI Tools