Introduction | 1 |
The Dawn of Alas a Consumer Product | 3 |
AI Shakes Up Entertainment | 7 |
Predictive AI: An Extraordinary Claim That Requires Extraordinary Evidence | 9 |
Painting AI with a Single Brush Is Tempting but Flawed | 12 |
A Series of Curious Circumstances Led to This Book | 18 |
The AI Hype Vortex | 21 |
What Is AI Snake Oil? | 26 |
Who This Book Is For | 34 |
How Predictive AI Goes Wrong | 36 |
Predictive AI Makes Life-Altering Decisions | 38 |
A Good Prediction Is Not a Good Decision | 43 |
Opaque AI Incentivizes Gaming | 46 |
Overautomation | 48 |
Predictions about the Wrong People | 51 |
Predictive AI Exacerbates Existing Inequalities | 53 |
A World without Prediction | 56 |
Concluding Thoughts | 58 |
Why Can't AI Predict the Future? | 60 |
A Brief History of Predicting the Future Using Computers | 62 |
Getting Specific | 67 |
The Fragile Families Challenge | 70 |
Why Did the Fragile Families Challenge End in Disappointment? | 73 |
Predictions in Criminal Justice | 78 |
Failure Is Hard. What about Success? | 81 |
The Meme Lottery | 86 |
From Individuals to Aggregates | 90 |
Recap: Reasons for Limits to Prediction | 97 |
The Long Road to Generative AI | 99 |
Generative AI Is Built on a Long Series of Innovations Dating Back Eighty Years | 105 |
Failure and Revival | 107 |
Training Machines to "See" | 111 |
The Technical and Cultural Significance of ImageNet | 114 |
Classifying and Generating Images | 118 |
Generative AI Appropriates Creative Labor | 122 |
AI for Image Classification Can Quickly Become AI for Surveillance | 127 |
From Images to Text | 129 |
From Models to Chatbots | 133 |
Automating Bullshit | 139 |
Deepfakes, Fraud, and Other Malicious Uses | 142 |
The Cost of Improvement | 143 |
Taking Stock | 146 |
Is Advanced AI an Existential Threat? | 150 |
What Do the Experts Think? | 151 |
The Ladder of Generality | 156 |
What's Next on the Ladder? | 162 |
Accelerating Progress ? | 165 |
Rogue AI? | 168 |
A Global Ban on Powerful AI? | 172 |
A Better Approach: Defending against Specific Threats | 174 |
Concluding Thoughts | 177 |
Why Can't AI Fix Social Media? | 179 |
When Everything Is Taken Out of Context | 183 |
Cultural Incompetence | 188 |
AI Excels at Predicting … the Past | 194 |
When AI Goes Up against Human Ingenuity | 198 |
A Matter of Life and Death | 201 |
Now Add Regulation into the Mix | 205 |
The Hard Part Is Drawing the Line | 209 |
Recap: Seven Shortcomings of AI for Content Moderation | 216 |
A Problem of Their Own Making | 218 |
The Future of Content Moderation | 223 |
Why Do Myths about AI Persist? | 227 |
AI Hype Is Different from Previous Technology Hype | 231 |
The AI Community Has a Culture and History of Hype | 235 |
Companies Have Few Incentives for Transparency | 239 |
The Reproducibility Crisis in AI Research | 241 |
News Media Misleads the Public | 247 |
Public Figures Spread AI Hype | 251 |
Cognitive Biases Lead Us Astray | 255 |
Where Do We Go from Here? | 258 |
AI Snake Oil Is Appealing to Broken Institutions | 261 |
Embracing Randomness | 265 |
Regulation: Cutting through the False Dichotomy | 268 |
Limitations of Regulation | 274 |
AI and the Future of Work | 276 |
Growing Up with AI in Kai's World | 281 |
Growing Up with AI in Maya's World | 285 |
Acknowledgments | 291 |
References | 293 |
Index | 331 |
Table of Contents provided by Ingram Syndetics. All Rights Reserved. |