Watch a provocative keynote address from USENIX Security '18 that challenges the notion that improving computer security is possible. Explore why the proliferation of ubiquitous technology may not be beneficial, using case studies in machine learning and other rapidly developed technologies. Examine the difficulties in achieving ethical computing and computer security when developers prioritize speed over introspection. Delve into topics such as AI mission creep, the pitfalls of machine learning in criminal justice, and the security risks of the Internet of Things. Gain insights into the speaker's skeptical view of technological progress and his call for a more thoughtful approach to development in the tech industry.
Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible
Overview
Syllabus
Introduction
Keynote Definition
Value of Skepticism
Machine Learning
Egg Drop
Other Winners Curse
Two Mistakes
Kingsley
AI Mission Creep
AI Workflow
Conversation
Tay
AI Child
Internet People
Twitter Answers
Why Is That
Race War
Grammar Lessons
Prejudice Science
Sotae Rant
A Holistic View of Security
Cartesian Security
ARPANET
Cartesian Approach
Gadgets Are Eternal
The Liberal Arts Are Getting Upset
Theyve Made Mistakes Too
Machine Learning In The Criminal Justice System
Gender Detection And Faces
White People Have Faces
Flashlight Apps
Technological Manifest Destiny
Statistics
Internet of Things
IOT Security
TLS
Cheapness
Best Seller
Have I written a paper about IoT
Why wouldnt you do this
Avoidable Mistakes
Conclusion
Questions
Would you use machine learning
Taught by
USENIX
