Index
action, in decision-making, 86, 124, 126, 127
Adobe, 214
adoption, timing of, 17, 182–184, 206
adversarial machine learning, 211–212
gender discrimination in, 222–224
agriculture
monoculture in, 227
AI. See artificial intelligence (AI)
AI-first strategy, 203–206, 218
AI Insight, 14
Air France Flight 447, 216–217
airline industry, 192–195, 216, 241–242
airline pilots, 208–209, 216–217
AI tools. See tools, AI
AI winter, 46
Alipay, 245
AliveCor, 58
Alphabet, 188
AlphaGo Zero, 248
AI asset acquisition by, 243
business model, 16–17, 180–181
fulfillment processes at, 167, 168, 180–181
low-stakes transactions at, 135–136, 138–140
Machine Learning, 229
Picking Challenge, 168
privacy and, 214
privacy policy, 246
reviews, 136
robots used by, 117
scale of, 241
American Express, 96
The Americans, 115
anticipatory shipping, 16–17, 180–181
Apple, 2
Siri, 157
artificial general intelligence (AGI), 157, 248–249
artificial intelligence (AI)
automation versus, 124
biases in, 48, 69–69, 71, 221–224
business transformation and, 191–202
decision-making and, 149
diversity in machines for, 228
economics of, 3, 8–10, 12–13, 189
as existential threat, 247–249
learning from mistakes by, 109
research and development of, 243–244
role of data in, 57
societal impacts of, 19, 235–250
strategy and, 16–17, 18–19, 20, 179–190
artificial light, 11
artistic achievement, 129
Asimov, Isaac, 127
AT&T, 241
athletic competition, 129
automatic teller machines (ATMs), 195–196
automation
AI versus, 124
legal requirements for humans with, 127–129
automobile industry, 168–169, 188–189, 193–194, 195. See also autonomous vehicles
autonomous delivery systems, 117
autonomous vehicles, 90
human deskilling and, 217
learning by, 213
legal requirements on, 128
passengers and, 107
prediction and, 14–15, 100–101
preferences and, 101
rail systems, 116
reward function engineering, 104
school bus drivers and, 173
security risks, 228
stakes of, 143
tolerance for error in, 143, 209–211
trolley problem and, 128
average, conditional, 47
back propagation, 52
bail-granting decisions, 70–71
baseball players, selection of, 69, 185
Bayesian estimation, 13
Beijing Automotive Group, 188–189
Bernoulli numbers, 12
beta testing, 208, 210–211, 216
Bhalla, Ajay, 25
biases, 19
big companies, control by, 241–243
Blake, Thomas, 225
blockchain, 246
Blum, Andrew, 38
boundary shifting, 181, 189–192
what to leave in/leave out, 192–194
brainpower, 239
Breakout, 207
Bresnahan, Tim, 12
A Brief History of Time (Hawking), 236–237
business models, 11, 16–17, 96, 180–181
business strategy. See strategy
business transformation, 191–202
Camelyon Grand Challenge, 78
Capital in the Twenty-First Century (Piketty), 239
Capital One, 111
carbon tax, 129
Cardiio, 58
reverse, 75
CDL. See Creative Destruction Lab (CDL)
certainty, 32
Challenger, 167
chatbots, 133
Chavez, R. Martin, 149
chess, 76
chief executive officer (CEO), AI adoption and, 180
China
autonomous driving in, 188–189
language translation in, 26
Chiou, Lesley, 242
Christensen, Clay, 205
classification, 13
climate change, 129
cloud
untethering from, 228
clustering, 13
coffeehouses, 31
cognitive costs, of judgment, 99–100
collaboration, human/machine
bank tellers/ATMs and, 195–196
collateralized debt obligations (CDOs), 51
college recruitment, 151–153, 162–163
competition, 206, 238–239, 241, 247, 248
complements, 15. See also data; judgment
complex environments, 52
computer revolution, 165
programming, 54
as tools, 157
conditional average, 47
consumer data, 109–110, 213–215, 245–247
consumer interactions, stakes of, 133–135
consumer preferences, data on, 200–201
consumer satisfaction, 205
cookies, 199
unanticipated, 51
cortex, 53
cost
effects of reduced AI, 9–11, 12–13
counterfactuals, 75–76, 110–111
creative destruction, 241
Creative Destruction Lab (CDL), 2, 8, 158
credit card fraud prevention, 24–25, 27, 96–99, 103
creditworthiness, 27, 79–80, 96
Croesus, King of Lydia, 23
crystal balls, 24
customer data, 109–110, 213–215, 245–247
customer service, 3
Daimler, 188
Dartmouth College, 45
in churn prediction, 50
consumer, 109–110, 213–215, 245–247
cost of acquiring, 58
crashes, 226
feedback, 57, 60, 86, 187, 230–231
input, 57, 58, 61, 86, 187, 226–228
machine learning and, 49–50, 59–61, 210–211
ownership and control of, 198–201
for prediction, 57–59, 109–110, 198–201, 226
real world, 210
roles of, in AI, 57
search engines and, 242
as strategic asset, 187
training, 57, 60, 61–62, 86, 187, 210, 228–230
unique, 200
value of, 57
AI and, 149
big and small, 85
certainty in, 32
experiments and, 111
judgment in, 86–88, 90–93, 95–106, 108, 160, 197–198
prediction and, 2, 4, 15, 18, 29
subjective goals and, 197
uncertainty in, 3, 32–35, 86, 90–93
See also uncertainty
Deep Genomics, 3
back propagation in, 52
flexibility in, 50
language translation and, 25–27
Deep Thinking (Kasparov), 76
delivery robots, 117
Delphi oracle, 23
dependent variables, 59
DePodesta, Paul, 185
deskilling, of humans, 216–218
deterministic programming, 54
digital cameras, 14
disruptive innovation, 179, 205–206
diversity, of prediction machines, 228
human/machine collaboration, 78–80
human weaknesses in prediction and, 68–71
machine weaknesses in prediction and, 71–78
prediction by exception and, 80–81
“dog fooding,” 208
drone weapons, 128
Dropbox, 214
Dubé, J. P., 105
D-Wave, 212
eBay, 225
economies of scale, 63–64, 241–243
Edelman, Ben, 222
education, 240
emergency braking, 124
employment. See jobs
Enlitic, 170
errors
consequences of, 143
ethical dilemmas, 128
Etzioni, Oren, 246
executive leadership, AI and, 179–190
experiments, 111
experts, predictions by, 68–69
dominance of social media, 241
facial recognition, 73, 109, 214–215, 245
false negatives, 136, 138, 139, 143
false positives, 136, 138, 139, 143
Fanning, Shawn, 74
Farmers’ Almanac, 37
Federal Trade Commission, 221
feedback data, 57, 60, 86, 187
films, about AI, 8
financial crisis (2008), 50–51
Forbes, Silke, 192
fraud detection, 24–25, 27, 96–99, 103
free-riding effect, 246
free trade, 239
Frey, Carl, 173
function engineering, 238
Furman, Jason, 239
Gates, Bill, 187, 236, 239, 247
gender discrimination, 222–224
general adversarial networks, 13
Gildert, Suzanne, 169
Go (game), 8
Goizueta, Robert, 57
Goldin, Claudia, 240
Goldman Sachs, 149
goodness of fit, 47
advertising, 200, 221–222, 224–225
AI development by, 184
AI-first strategy, 203–204, 205, 218
AI tool development by, 150
algorithm of, 77
anti-spam team, 229
data of, 57
search engine, 10, 19, 64, 157, 229, 241
Waymo and, 107
Google Assistant, 118
government expenditure, on AI, 243–244
Grammarly, 108
Griliches, Zvi, 183
Grove, Andy, 179
Hacking, Ian, 54
Harford, Tim, 217
Hawkins, Jeff, 53
Hinton, Geoffrey, 169
hiring decisions, 71
ZipRecruiter and, 104–106, 111–112
Hoffman, Mitchell, 71
Houston Astros, 185
human resources (HR) management, 197
human senses, 109
Hume, Kathryn, 14
humor, 129
IBM, 170
identity verification, 227
iFlytek, 26
if-then logic, 14, 103, 116–121
ImageNet, 7
incumbents, 206
independent variables, 59
initial public offerings (IPOs), 10, 149
innovation, 193
innovator’s dilemma, 206
input data, 57, 58, 61, 86, 187
Instagram, 77
intelligence
See also artificial intelligence (AI)
intuition, 87
inventory management, 28, 181–182
iPhone
introduction of, 179
James, Bill, 69
Jelinek, Frederick, 120
jobs
AI’s impact on, 195–198, 236–238, 239
augmentation of, 166
impact of computers on, 165–166
See also workflows
jokes, 129
judgment, 18
in credit card fraud, 96–99, 103
in decision-making, 86–88, 90–93, 95–106, 108, 160, 197–198
jobs related to, 238
preferences and, 100–102, 108–109
in reward function engineering, 103–106
strategic choice and, 186
uncertainty and, 103
Juhong, Chen, 188
Kahn, Lisa, 71
Kahneman, Daniel, 68, 69, 235–236
Kapor, Mitch, 187
Kasparov, Garry, 76
Katz, Lawrence, 240
Kie, Ke, 8
Kindred, 169
Kiva, 167
known unknowns, 71, 72–73, 78, 110
Kurzweil, Ray, 247
See also jobs
“Lady Lovelace’s Objection,” 13
language translation, 25–27, 119–120
learning
in the cloud versus on the ground, 212–213, 228
reinforcement learning, 13, 207–208
supervised, 207
Lederman, Mara, 192
Lee, Kai-Fu, 245
legal issues, with automation, 127–129
Lewis, Michael, 69
Li, Danielle, 71
lighting, cost of, 11
Lloyd, Edward, 31
Lloyd’s List, 31
Lloyd’s of London, 33
loyalty cards, 110
Lu, Qi, 245
Lyft, 100
Lytvyn, Max, 108
machine learning, 2
as artificial intelligence, 51–54
data use by, 49–50, 59–61, 210–211
experiments in, 49
flexibility in, 50
one-shot, 73
pattern recognition and, 169–170
quantum, 212
regression compared with, 46–50
statistics, prediction, and, 51–54
undermining, by bad actors, 230–231
management by exception, 81
Mastercard, 96
Mazda, 148
MBA program recruitment, 151–153, 162–163
medical diagnosis, 120–121, 169–172, 191–192, 201
medicine, 3
Mejdal, Sig, 185
Microsoft, 9, 200, 204, 205, 218, 241, 245
mining, automation in, 124–126
Misra, Sanjog, 105
mistakes, learning from, 109
mobile-first strategy, 204
Mobileye, 15
monoculture, 227
monopolies, 241
multivariate regression, 47
music, 12
music industry, 74
Napster, 74
NASA, 2
national income, 239
National Science and Technology Council (NSTC), 248–249
navigation apps, 89–90, 100–101, 215
Neilley, Peter, 39
neocortex, 53
Netflix, 133
neural networks, 13
New Economy, 10
New York City Fire Department, 223
Ng, Andrew, 240
Nordhaus, William, 11
Norvig, Peter, 204
Nosko, Chris, 225
Novak, Sharon, 193
Nymi, 227
object identification, 28–29, 52, 169–170
object recognition, 7
omitted variables, 75
one-shot learning, 73
On Intelligence (Hawkins), 53
operational performance, 205
option value, 40
organic light-emitting diode (OLED), 212
Osborne, Michael, 173
OTI Lumionics, 212
Oura, 58
Page, Larry, 203
Paravisini, Daniel, 79
Pavlov, Ivan, 207
payoffs, 86, 87–88, 92–93, 99–100, 160
Pell, Barney, 2
performance evaluation, 197
personalization, 140–142, 245–246
pharmaceutical industry, 158–162
photography, 14
physical stores, 181
Piketty, Thomas, 239
power calculations, 62
about, 13
AI and, 13
in the cloud versus on the ground, 228
consequences of cheap, 29, 53–54
credit card fraud prevention and, 24–25
data for, 57–59, 86, 109–110, 198–201, 226
decision-making and, 2, 4, 15, 18
definition of, 24
human/machine collaboration for, 78–80
imperfect, 134
improvements in accuracy of, 27–29
language translation and, 25–27
for risk mitigation, 35–36, 40–42
statistics and, 68
techniques, 13
trade-offs, 20
uncertainty and, 3
presidential elections, 72–73, 110
privacy
data, 4, 63, 110, 213–215, 245–246
search engine, 64
probabilities, 171
probabilistic programming, 54
processes. See workflows
product personalization, 140–142
Project Apollo, 188
protected classes, 223
protection, 135
protection strategy, to mitigate risk, 40–42
Puchai, Sundar, 203
Putin, Vladimir, 243
quantum computing, 212
rail systems, 116
randomized control trials, 111
reason, 53
recruitment, to MBA programs, 151–153, 162–163
regional airlines, 192–193, 194
multivariate, 47
reinforcement learning, 13, 169, 207–208
research and development, 243–244
reverse causality, 75
reward function, 91
reward function engineering, 103–106
Rio Tinto, 125
risk aversion, 32
risk mitigation, 32
risks and risk management, 19, 40–42, 221–232
prediction and reducing, 120–121
Rivers, Lynn, 244
Robotlandia, 237
in space, 127
Roomba, 116
Rose, Geordie, 169
Russia, 243
scale economies, 63–64, 241–243
Schoar, Antoinette, 79
school bus drivers, 173
Schumpeter, Joseph, 241
science-based startups, 2
search engine optimization, 77
search engines, 242
Se-dol, Lee, 8
self-driving cars. See autonomous vehicles
Shevchenko, Alex, 108
Simon, Herbert, 119
skills
need for additional, 240
Skynet, 53
control of big companies and, 241–243
country advantages and, 243–247
existential threat to, 247–249
impacts of AI on, 19
Solow, Robert, 147
space exploration, 127
spammers, 77
sports
camera automation and, 126–127
Sputnik, 8
of consumer interactions, 133–135
statistics
prediction, machine learning, and, 51–54
prediction and, 68
stereotypes, 19
Stigler, George, 117
strategic choice, 186
AI’s impact on, 18–20, 179–190
business transformation and, 191–202
economics of, 189
judgment and, 186
organizational structure and, 185–186
strong AI, 157
Sullenberg, Chesley “Sully,” 208
supervised learning, 207
Sweeney, Latanya, 221
Tadelis, Steve, 225
Taming of Chance (Hacking), 54
Tanner, Adam, 221
task analysis, job redesign and, 166–169, 173–174
taxi industry, 179
technical support services, 102
technological change, price reductions and, 11–12
technology adoption, timing of, 17, 182–184, 206
technophiles, 7
technophobes, 7
telecommunications, 241
Teradata Center, 49
Tesla, 8, 101, 123–124, 126, 128, 133–135, 157, 189, 210–211, 213
Thinking, Fast and Slow (Kahneman), 235–236
Tinder, 213
adoption of, 184
decisions about using, 157–164
for deconstructing workflows, 147–155
design of, 149
development of, 184
impact on workflows of, 150–153
topological data analysis, 13
trade-offs, 4, 5, 20, 193–194, 204
between human and machine experience, 217–218
of when to deploy, 209
traffic apps, 118
training
training data, 57, 60, 61–62, 86, 187, 210
transportation, 14
trolley problem, 128
true negatives, 136
true positives, 136
Tucker, Catherine, 222, 224, 242
Tunstall-Pedoe, William, 2
Turing, Alan, 13
Turing Award, 119
Turing test, 53
Tversky, Amos, 68
2001: A Space Odyssey, 247
business boundaries and, 192–194
control and, 193
data to reduce, 57
in decision-making, 32, 33–35, 86, 90–93
increases in, 198
judgment and, 103
underwriters, 33
unemployment rates, 237
unknown knowns, 72, 74–78, 110–111, 225
unknown unknowns, 71–72, 73–74, 109, 113
Validere, 3
variables
dependent, 59
independent, 59
omitted, 75
Varian, Hal, 57
Vinge, Vernor, 247
Visa, 96
Wanamaker, John, 199
War Games, 143
Watson, 170
weather forecasts, 35–39, 192–193
web advertising, 199–200, 221–225
WeChat, 188
Wells Fargo, 197
Windows 95, 9
workflows
impact of AI tools on, 150–153
redesigning for automation, 147–150
World War II, 35–37, 112–113, 142
WTF problem, 141
Xanadu, 212
Yahoo, 242
Y Combinator, 236
Yeomans, Mike, 129
YouTube, 200