AI Essentials For Business - Module 3 Lesson 2 - Ethics of Digital Scale Scope and Learning
AI Essentials For Business - Module 3 Lesson 2 - Ethics of Digital Scale Scope and Learning
Ethics of Digital
Scale, Scope, and Learning
Part 1
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Ethics of Digital
Scale, Scope, and Learning
Part 2
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Privacy Inclusiveness
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Digital Amplification
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Echo Chambers
Graph shows democratic-leaning and republican-leaning internet posts and reposts; the few
dots in the middle show examples of reposts across the political divide.
Recreated by Liette Ocker based on image and information from: Kiran Garimella, Gianmarco De Francisci
Morales, Aristides Gionis, and Michael Mathioudakis. Reducing opposing views to reduce controversy.
Proceedings of the 10th ACM International Conference on Web Search and Data Mining, 2017.
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Algorithmic Bias
Introduction
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Apple Card
Bias
Algorithmic Justice League Chart
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Sources of bias
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Cybersecurity
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Cybersecurity
Cybercrime to cost the world $10.5 trillion annually by 2025
Recreated by Liette Ocker based on AFP image and information from BusinessInsider.com
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Cybersecurity (cont.)
FBI Internet Crime Complaint Center (IC3) 2019 annual report
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Privacy
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Privacy
• Privacy and cybersecurity becoming more integrated
• More countries and U.S. states will introduce new data
protection laws
• Privacy becoming much greater focus in board-level
discussions
• More companies moving to enterprise-wide privacy
strategy
• Emergence of privacy technologies
(e.g., differential privacy), but shortage of experts
• Privacy becoming a business differentiator
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Differential Privacy
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Differential Privacy
• Anonymizing data is not enough to protect personal information.
• With differential privacy, the goal is to give everyone roughly the same
privacy that would result from having their data removed (without removing
the data), for a given number of hostile queries. We want to gain some
utility from our database, without exposing any sensitive data to a possible
adversary.
This is done by adding noise to the dataset until we achieve the desired privacy guarantee.
Adapted from https://georgianpartners.shinyapps.io/interactive_counting/
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Inclusiveness
HARVARD BUSINESS SCHOOL
COMPETING IN THE AGE OF AI PROGRAM
Inclusiveness