-
Notifications
You must be signed in to change notification settings - Fork 0
/
ai-policy-and-u-s-china-relations.html
402 lines (351 loc) · 40.1 KB
/
ai-policy-and-u-s-china-relations.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml" lang="" xml:lang="">
<head>
<meta charset="utf-8" />
<meta http-equiv="X-UA-Compatible" content="IE=edge" />
<title>4 AI policy and U.S.-China relations | Artificial Intelligence: American Attitudes and Trends</title>
<meta name="description" content="4 AI policy and U.S.-China relations | Artificial Intelligence: American Attitudes and Trends" />
<meta name="generator" content="bookdown 0.12 and GitBook 2.6.7" />
<meta property="og:title" content="4 AI policy and U.S.-China relations | Artificial Intelligence: American Attitudes and Trends" />
<meta property="og:type" content="book" />
<meta name="twitter:card" content="summary" />
<meta name="twitter:title" content="4 AI policy and U.S.-China relations | Artificial Intelligence: American Attitudes and Trends" />
<meta name="author" content="Baobao Zhang and Allan Dafoe" />
<meta name="author" content="Center for the Governance of AI, Future of Humanity Institute, University of Oxford" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<meta name="apple-mobile-web-app-capable" content="yes" />
<meta name="apple-mobile-web-app-status-bar-style" content="black" />
<link rel="prev" href="public-opinion-on-ai-governance.html">
<link rel="next" href="trend-across-time-attitudes-toward-workplace-automation.html">
<script src="libs/jquery-2.2.3/jquery.min.js"></script>
<link href="libs/gitbook-2.6.7/css/style.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-table.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-bookdown.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-highlight.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-search.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-fontsettings.css" rel="stylesheet" />
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-132060565-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-132060565-1');
</script>
</head>
<body>
<div class="book without-animation with-summary font-size-2 font-family-1" data-basepath=".">
<div class="book-summary">
<nav role="navigation">
<ul class="summary">
<li><img src="images/small_logo.png" alt="Report small logo" width="272px" hspace="12" vspace="12"/></li>
<li><a href="index.html"><b>Table of Contents</b></a></li>
<li class="divider"></li>
<li class="chapter" data-level="1" data-path="executive-summary.html"><a href="executive-summary.html"><i class="fa fa-check"></i><b>1</b> Executive summary</a><ul>
<li class="chapter" data-level="1.1" data-path="executive-summary.html"><a href="executive-summary.html#select-results"><i class="fa fa-check"></i><b>1.1</b> Select results</a></li>
<li class="chapter" data-level="1.2" data-path="executive-summary.html"><a href="executive-summary.html#reading-notes"><i class="fa fa-check"></i><b>1.2</b> Reading notes</a></li>
<li class="chapter" data-level="1.3" data-path="executive-summary.html"><a href="executive-summary.html#press-coverage"><i class="fa fa-check"></i><b>1.3</b> Press coverage</a></li>
</ul></li>
<li class="chapter" data-level="2" data-path="general-attitudes-toward-ai.html"><a href="general-attitudes-toward-ai.html"><i class="fa fa-check"></i><b>2</b> General attitudes toward AI</a><ul>
<li class="chapter" data-level="2.1" data-path="general-attitudes-toward-ai.html"><a href="general-attitudes-toward-ai.html#subsecsupportai"><i class="fa fa-check"></i><b>2.1</b> More Americans support than oppose developing AI</a></li>
<li class="chapter" data-level="2.2" data-path="general-attitudes-toward-ai.html"><a href="general-attitudes-toward-ai.html#subsecdemosupportai"><i class="fa fa-check"></i><b>2.2</b> Support for developing AI is greater among those who are wealthy, educated, male, or have experience with technology</a></li>
<li class="chapter" data-level="2.3" data-path="general-attitudes-toward-ai.html"><a href="general-attitudes-toward-ai.html#subsecsupportmanageai"><i class="fa fa-check"></i><b>2.3</b> An overwhelming majority of Americans think that AI and robots should be carefully managed</a></li>
<li class="chapter" data-level="2.4" data-path="general-attitudes-toward-ai.html"><a href="general-attitudes-toward-ai.html#harmful-consequences-of-ai-in-the-context-of-other-global-risks"><i class="fa fa-check"></i><b>2.4</b> Harmful consequences of AI in the context of other global risks</a></li>
<li class="chapter" data-level="2.5" data-path="general-attitudes-toward-ai.html"><a href="general-attitudes-toward-ai.html#americans-understanding-of-key-technology-terms"><i class="fa fa-check"></i><b>2.5</b> Americans’ understanding of key technology terms</a></li>
</ul></li>
<li class="chapter" data-level="3" data-path="public-opinion-on-ai-governance.html"><a href="public-opinion-on-ai-governance.html"><i class="fa fa-check"></i><b>3</b> Public opinion on AI governance</a><ul>
<li class="chapter" data-level="3.1" data-path="public-opinion-on-ai-governance.html"><a href="public-opinion-on-ai-governance.html#subsecgovchallenges13"><i class="fa fa-check"></i><b>3.1</b> Americans consider many AI governance challenges to be important; prioritize data privacy and preventing AI-enhanced cyber attacks, surveillance, and digital manipulation</a></li>
<li class="chapter" data-level="3.2" data-path="public-opinion-on-ai-governance.html"><a href="public-opinion-on-ai-governance.html#americans-who-are-younger-who-have-cs-or-engineering-degrees-express-less-concern-about-ai-governance-challenges"><i class="fa fa-check"></i><b>3.2</b> Americans who are younger, who have CS or engineering degrees express less concern about AI governance challenges</a></li>
<li class="chapter" data-level="3.3" data-path="public-opinion-on-ai-governance.html"><a href="public-opinion-on-ai-governance.html#americans-place-the-most-trust-in-the-u.s.-military-and-universities-to-build-ai-trust-tech-companies-and-non-governmental-organizations-more-than-the-government-to-manage-the-technology"><i class="fa fa-check"></i><b>3.3</b> Americans place the most trust in the U.S. military and universities to build AI; trust tech companies and non-governmental organizations more than the government to manage the technology</a></li>
</ul></li>
<li class="chapter" data-level="4" data-path="ai-policy-and-u-s-china-relations.html"><a href="ai-policy-and-u-s-china-relations.html"><i class="fa fa-check"></i><b>4</b> AI policy and U.S.-China relations</a><ul>
<li class="chapter" data-level="4.1" data-path="ai-policy-and-u-s-china-relations.html"><a href="ai-policy-and-u-s-china-relations.html#americans-underestimate-the-u.s.-and-chinas-ai-research-and-development"><i class="fa fa-check"></i><b>4.1</b> Americans underestimate the U.S. and China’s AI research and development</a></li>
<li class="chapter" data-level="4.2" data-path="ai-policy-and-u-s-china-relations.html"><a href="ai-policy-and-u-s-china-relations.html#subsecexperimentchina"><i class="fa fa-check"></i><b>4.2</b> Communicating the dangers of a U.S.-China arms race requires explaining policy trade-offs</a></li>
<li class="chapter" data-level="4.3" data-path="ai-policy-and-u-s-china-relations.html"><a href="ai-policy-and-u-s-china-relations.html#americans-see-the-potential-for-u.s.-china-cooperation-on-some-ai-governance-challenges"><i class="fa fa-check"></i><b>4.3</b> Americans see the potential for U.S.-China cooperation on some AI governance challenges</a></li>
</ul></li>
<li class="chapter" data-level="5" data-path="trend-across-time-attitudes-toward-workplace-automation.html"><a href="trend-across-time-attitudes-toward-workplace-automation.html"><i class="fa fa-check"></i><b>5</b> Trend across time: attitudes toward workplace automation</a><ul>
<li class="chapter" data-level="5.1" data-path="trend-across-time-attitudes-toward-workplace-automation.html"><a href="trend-across-time-attitudes-toward-workplace-automation.html#americans-do-not-think-that-labor-market-disruptions-will-increase-with-time"><i class="fa fa-check"></i><b>5.1</b> Americans do not think that labor market disruptions will increase with time</a></li>
<li class="chapter" data-level="5.2" data-path="trend-across-time-attitudes-toward-workplace-automation.html"><a href="trend-across-time-attitudes-toward-workplace-automation.html#extending-the-historical-time-trend"><i class="fa fa-check"></i><b>5.2</b> Extending the historical time trend</a></li>
</ul></li>
<li class="chapter" data-level="6" data-path="high-level-machine-intelligence.html"><a href="high-level-machine-intelligence.html"><i class="fa fa-check"></i><b>6</b> High-level machine intelligence</a><ul>
<li class="chapter" data-level="6.1" data-path="high-level-machine-intelligence.html"><a href="high-level-machine-intelligence.html#arrivesooner"><i class="fa fa-check"></i><b>6.1</b> The public predicts a 54% likelihood of high-level machine intelligence within 10 years</a></li>
<li class="chapter" data-level="6.2" data-path="high-level-machine-intelligence.html"><a href="high-level-machine-intelligence.html#subsecsupporthlmi"><i class="fa fa-check"></i><b>6.2</b> Americans express mixed support for developing high-level machine intelligence</a></li>
<li class="chapter" data-level="6.3" data-path="high-level-machine-intelligence.html"><a href="high-level-machine-intelligence.html#subsecdemohlmi"><i class="fa fa-check"></i><b>6.3</b> High-income Americans, men, and those with tech experience express greater support for high-level machine intelligence</a></li>
<li class="chapter" data-level="6.4" data-path="high-level-machine-intelligence.html"><a href="high-level-machine-intelligence.html#subsecharmgood"><i class="fa fa-check"></i><b>6.4</b> The public expects high-level machine intelligence to be more harmful than good</a></li>
</ul></li>
<li class="appendix"><span><b>Appendices</b></span></li>
<li class="chapter" data-level="A" data-path="appmethod.html"><a href="appmethod.html"><i class="fa fa-check"></i><b>A</b> Appendix A: Methodology</a><ul>
<li class="chapter" data-level="A.1" data-path="appmethod.html"><a href="appmethod.html#yougovsampling"><i class="fa fa-check"></i><b>A.1</b> YouGov sampling and weights</a></li>
<li class="chapter" data-level="A.2" data-path="appmethod.html"><a href="appmethod.html#appdemosubgroups"><i class="fa fa-check"></i><b>A.2</b> Demographic subgroups</a></li>
<li class="chapter" data-level="A.3" data-path="appmethod.html"><a href="appmethod.html#appanalysis"><i class="fa fa-check"></i><b>A.3</b> Analysis</a></li>
<li class="chapter" data-level="A.4" data-path="appmethod.html"><a href="appmethod.html#datasharing"><i class="fa fa-check"></i><b>A.4</b> Data sharing</a></li>
</ul></li>
<li class="chapter" data-level="B" data-path="apptopline.html"><a href="apptopline.html"><i class="fa fa-check"></i><b>B</b> Appendix B: Topline questionnaire</a><ul>
<li class="chapter" data-level="B.1" data-path="apptopline.html"><a href="apptopline.html#global_risks"><i class="fa fa-check"></i><b>B.1</b> Global risks</a></li>
<li class="chapter" data-level="B.2" data-path="apptopline.html"><a href="apptopline.html#considersai"><i class="fa fa-check"></i><b>B.2</b> Survey experiment: what the public considers AI, automation, machine learning, and robotics</a></li>
<li class="chapter" data-level="B.3" data-path="apptopline.html"><a href="apptopline.html#knowledge-of-computer-science-cstechnology"><i class="fa fa-check"></i><b>B.3</b> Knowledge of computer science (CS)/technology</a></li>
<li class="chapter" data-level="B.4" data-path="apptopline.html"><a href="apptopline.html#supportdevai"><i class="fa fa-check"></i><b>B.4</b> Support for developing AI</a></li>
<li class="chapter" data-level="B.5" data-path="apptopline.html"><a href="apptopline.html#manageexp"><i class="fa fa-check"></i><b>B.5</b> Survey experiment: AI and/or robots should be carefully managed</a></li>
<li class="chapter" data-level="B.6" data-path="apptopline.html"><a href="apptopline.html#trustdevai"><i class="fa fa-check"></i><b>B.6</b> Trust of actors to develop AI</a></li>
<li class="chapter" data-level="B.7" data-path="apptopline.html"><a href="apptopline.html#trustmanageai"><i class="fa fa-check"></i><b>B.7</b> Trust of actors to manage AI</a></li>
<li class="chapter" data-level="B.8" data-path="apptopline.html"><a href="apptopline.html#govchallenges"><i class="fa fa-check"></i><b>B.8</b> AI governance challenges</a></li>
<li class="chapter" data-level="B.9" data-path="apptopline.html"><a href="apptopline.html#airesearchcompare"><i class="fa fa-check"></i><b>B.9</b> Survey experiment: comparing perceptions of U.S. vs. China AI research and development</a></li>
<li class="chapter" data-level="B.10" data-path="apptopline.html"><a href="apptopline.html#armsraceexp"><i class="fa fa-check"></i><b>B.10</b> Survey experiment: U.S.-China arms race</a><ul>
<li class="chapter" data-level="B.10.1" data-path="apptopline.html"><a href="apptopline.html#control"><i class="fa fa-check"></i><b>B.10.1</b> Control</a></li>
<li class="chapter" data-level="B.10.2" data-path="apptopline.html"><a href="apptopline.html#nationalism-treatment"><i class="fa fa-check"></i><b>B.10.2</b> Nationalism treatment</a></li>
<li class="chapter" data-level="B.10.3" data-path="apptopline.html"><a href="apptopline.html#war-risks-treatment"><i class="fa fa-check"></i><b>B.10.3</b> War risks treatment</a></li>
<li class="chapter" data-level="B.10.4" data-path="apptopline.html"><a href="apptopline.html#common-humanity-treatment"><i class="fa fa-check"></i><b>B.10.4</b> Common humanity treatment</a></li>
</ul></li>
<li class="chapter" data-level="B.11" data-path="apptopline.html"><a href="apptopline.html#uschinacoop"><i class="fa fa-check"></i><b>B.11</b> Issue areas for possible U.S.-China cooperation</a></li>
<li class="chapter" data-level="B.12" data-path="apptopline.html"><a href="apptopline.html#jobtime"><i class="fa fa-check"></i><b>B.12</b> Trend across time: job creation or job loss</a></li>
<li class="chapter" data-level="B.13" data-path="apptopline.html"><a href="apptopline.html#forecasthlmi"><i class="fa fa-check"></i><b>B.13</b> High-level machine intelligence: forecasting timeline</a></li>
<li class="chapter" data-level="B.14" data-path="apptopline.html"><a href="apptopline.html#supporthlmi"><i class="fa fa-check"></i><b>B.14</b> Support for developing high-level machine intelligence</a></li>
<li class="chapter" data-level="B.15" data-path="apptopline.html"><a href="apptopline.html#expectedoutcome"><i class="fa fa-check"></i><b>B.15</b> Expected outcome of high-level machine intelligence</a></li>
</ul></li>
<li class="chapter" data-level="C" data-path="addresults.html"><a href="addresults.html"><i class="fa fa-check"></i><b>C</b> Appendix C: Additional data analysis results</a><ul>
<li class="chapter" data-level="C.1" data-path="addresults.html"><a href="addresults.html#addsupportdevai"><i class="fa fa-check"></i><b>C.1</b> Support for developing AI</a></li>
<li class="chapter" data-level="C.2" data-path="addresults.html"><a href="addresults.html#addcarefullym"><i class="fa fa-check"></i><b>C.2</b> Survey experiment and cross-national comparison: AI and/or robots should be carefully managed</a></li>
<li class="chapter" data-level="C.3" data-path="addresults.html"><a href="addresults.html#appglobalrisks"><i class="fa fa-check"></i><b>C.3</b> Harmful consequences of AI in the context of other global risks</a></li>
<li class="chapter" data-level="C.4" data-path="addresults.html"><a href="addresults.html#aawhatsai"><i class="fa fa-check"></i><b>C.4</b> Survey experiment: what the public considers AI, automation, machine learning, and robotics</a></li>
<li class="chapter" data-level="C.5" data-path="addresults.html"><a href="addresults.html#appgovchallenges"><i class="fa fa-check"></i><b>C.5</b> AI governance challenges: prioritizing governance challenges</a></li>
<li class="chapter" data-level="C.6" data-path="addresults.html"><a href="addresults.html#trust-in-various-actors-to-develop-and-manage-ai-in-the-interest-of-the-public"><i class="fa fa-check"></i><b>C.6</b> Trust in various actors to develop and manage AI in the interest of the public</a></li>
<li class="chapter" data-level="C.7" data-path="addresults.html"><a href="addresults.html#appuschinacomp"><i class="fa fa-check"></i><b>C.7</b> Survey experiment: comparing perceptions of U.S. vs. China AI research and development</a></li>
<li class="chapter" data-level="C.8" data-path="addresults.html"><a href="addresults.html#appuschinaarmsrace"><i class="fa fa-check"></i><b>C.8</b> Survey experiment: U.S.-China arms race</a></li>
<li class="chapter" data-level="C.9" data-path="addresults.html"><a href="addresults.html#appjobloss"><i class="fa fa-check"></i><b>C.9</b> Trend across time: job creation or job loss</a></li>
<li class="chapter" data-level="C.10" data-path="addresults.html"><a href="addresults.html#apphlmi"><i class="fa fa-check"></i><b>C.10</b> High-level machine intelligence: forecasting timeline</a></li>
<li class="chapter" data-level="C.11" data-path="addresults.html"><a href="addresults.html#appsupporthlmi"><i class="fa fa-check"></i><b>C.11</b> Support for developing high-level machine intelligence</a></li>
<li class="chapter" data-level="C.12" data-path="addresults.html"><a href="addresults.html#appexpectedoutcome"><i class="fa fa-check"></i><b>C.12</b> Expected outcome of high-level machine intelligence</a></li>
</ul></li>
<li class="chapter" data-level="" data-path="acknowledgements.html"><a href="acknowledgements.html"><i class="fa fa-check"></i>Acknowledgements</a><ul>
<li class="chapter" data-level="" data-path="acknowledgements.html"><a href="acknowledgements.html#primary-researchers"><i class="fa fa-check"></i>Primary researchers</a></li>
<li class="chapter" data-level="" data-path="acknowledgements.html"><a href="acknowledgements.html#editing-and-design"><i class="fa fa-check"></i>Editing and design</a></li>
<li class="chapter" data-level="" data-path="acknowledgements.html"><a href="acknowledgements.html#funders"><i class="fa fa-check"></i>Funders</a></li>
<li class="chapter" data-level="" data-path="acknowledgements.html"><a href="acknowledgements.html#for-media-or-other-inquiries"><i class="fa fa-check"></i>For media or other inquiries</a></li>
<li class="chapter" data-level="" data-path="acknowledgements.html"><a href="acknowledgements.html#recommended-citation"><i class="fa fa-check"></i>Recommended citation</a></li>
</ul></li>
<li class="chapter" data-level="" data-path="about-us.html"><a href="about-us.html"><i class="fa fa-check"></i>About us</a><ul>
<li class="chapter" data-level="" data-path="about-us.html"><a href="about-us.html#about-the-center-for-the-governance-of-ai"><i class="fa fa-check"></i>About the Center for the Governance of AI</a></li>
<li class="chapter" data-level="" data-path="about-us.html"><a href="about-us.html#about-the-future-of-humanity-institute"><i class="fa fa-check"></i>About the Future of Humanity Institute</a></li>
</ul></li>
<li class="chapter" data-level="" data-path="references.html"><a href="references.html"><i class="fa fa-check"></i>References</a></li>
<li class="divider"></li>
<li><a href="https://governance.ai">Center for the Governance of AI</a></li>
<li><a href="https://www.fhi.ox.ac.uk/">Future of Humanity Institute</a></li>
<li><a href="http://www.ox.ac.uk/">University of Oxford</a></li>
<li><img src="images/FHI-Logo-Print.png" alt="FHI logo" width="77px" hspace="12"/><img src="images/oxford-university-logo.png" alt="Oxford logo" width="74px" hspace="12"/></li>
</ul>
</nav>
</div>
<div class="book-body">
<div class="body-inner">
<div class="book-header" role="navigation">
<h1>
<i class="fa fa-circle-o-notch fa-spin"></i><a href="./">Artificial Intelligence: American Attitudes and Trends</a>
</h1>
</div>
<div class="page-wrapper" tabindex="-1" role="main">
<div class="page-inner">
<section class="normal" id="section-">
<div id="ai-policy-and-u.s.-china-relations" class="section level1">
<h1><span class="header-section-number">4</span> AI policy and U.S.-China relations</h1>
<div id="americans-underestimate-the-u.s.-and-chinas-ai-research-and-development" class="section level2">
<h2><span class="header-section-number">4.1</span> Americans underestimate the U.S. and China’s AI research and development</h2>
<!--
BZ: I think this section is 85% there. This is a tricky topic to write about so I try to contextualize the results by providing background information. We should emphasize this is probably an ignorant perception on the part of the American public.
-->
<p>In this survey experiment, we asked respondents to consider either the U.S. or China’s status in AI research and development (R&D). (Details of the survey experiment are found in <a href="apptopline.html#airesearchcompare">Appendix B</a>.) Respondents were asked the following:</p>
<blockquote>
<p>Compared with other industrialized countries, how would you rate [the U.S./China] in AI research and development?</p>
</blockquote>
<p>By almost any metric of absolute achievement (not per-capita achievement), the U.S. and China are the world leaders in the research and development of AI. The U.S. and China led participation in the 2017 AAAI Conference, one of the important ones in the field of AI research; 34% of those who presented papers had a U.S. affiliation while 23% had a Chinese affiliation <span class="citation">(Goldfarb and Trefler <a href="#ref-goldfarb2018ai">2018</a>)</span>. The U.S. and China also have the highest percentage of the world’s AI companies, 42% and 23%, respectively <span class="citation">(IT Juzi and Tencent Institute <a href="#ref-chinausreport2017">2017</a>)</span>. Most clearly, the U.S. and China have the largest technology companies focused on developing and using AI (Google, Facebook, and Amazon in the U.S.; Tencent, Alibaba, and Baidu in China).</p>
<!--
AD: can you flesh out above. perhaps skim Jeff's China report. Then ask jeff ding for best metrics/reference on this. And then refine the statement.
Also, by many metrics of per-capita achievement, the U.S. and China are also world leaders in AI R&D.
[BZ: I filled in some...I will send the rest to Jeff. Why do you want to tease apart the absolute metric versus per-capita metric?]
BZ to JD: Can you please add in other important R&D metrics? Also, the Li 2017 URL link is broken. Can you send me a URL that works? Also, do you have per capita metrics?
-->
<p>Yet, only a minority of the American public thinks the U.S. or China’s AI R&D is the “best in the world,” as reported in Figure <a href="ai-policy-and-u-s-china-relations.html#fig:uschina">4.1</a>. Our survey result seems to reflect the gap between experts and the public’s perceptions of U.S.’s scientific achievements in general. While 45% of scientists in the American Association for the Advancement of Science think that scientific achievements in the U.S. are the best in the world, only 15% of the American public express the same opinion <span class="citation">(Funk and Rainie <a href="#ref-funk2015">2015</a>)</span>.</p>
<p>According to our survey, there is not a clear perception by Americans that the U.S. has the best AI R&D in the world. While 10% of Americans believe that the U.S. has the best AI R&D in the world, 7% think that China does. Still, 36% of Americans believe that the U.S.’s AI R&D is “above average” while 45% think China’s is “above average.” Combining these into a single measure of whether the country has “above average” or “best in the world” AI R&D, Americans do not perceive the U.S. to be superior, and the results lean towards the perception that China is superior. Note that we did not ask for a direct comparison, but instead asked each respondent to evaluate one country independently on an absolute scale <a href="addresults.html#appuschinacomp">Appendix C</a>.</p>
<!-- BZ 19-12: Since the different is only significant when we control for missing, I moved the result to this footnote instead.
AD 18-12: I removed this. I don't think the "after controlling for" business is worth putting in main text or footnote. It's a methods decision, which isn't clear to me, which we should just make our best call on (and declare in appendix).
What is meant by after controlling for that?
[^airdcompare]: According to our regression analysis in [Appendix C](#appuschinacomp), the respondents think China's R&D is slightly better than the U.S.'s. This result is only statistically significant at the 5% level after controlling for don't know or missing responses.
Old text:
Americans do not perceive the U.S. to be better at AI R&D than China, according to our survey. In fact, according to our [regression analysis](#appuschinacomp), the respondents think China is slightly better than the U.S. Ten percent of Americans believe that the U.S. is the best in the world regarding AI R&D, while 7% think that about China. In contrast, 36% of Americans believe that the U.S.'s AI R&D is "above average" while 45% think China's is "above average." Nevertheless, a quarter of the respondents indicated that they do not know the quality of the U.S. or China's R&D.
-->
<p>Our results mirror those from a recent survey that finds that Americans think that China’s AI capability will be on par with the U.S.’s in 10 years <span class="citation">(West <a href="#ref-west2018worries">2018</a><a href="#ref-west2018worries">b</a>)</span>. The American public’s perceptions could be caused by media narratives that China is catching up to the U.S. in AI capability <span class="citation">(Kai-Fu <a href="#ref-kai2018ai">2018</a>)</span>. Nevertheless, another study suggests that although China has greater access to big data than the U.S., China’s AI capability is about half of the U.S.’s <span class="citation">(Ding <a href="#ref-ding2018">2018</a>)</span>. Exaggerating China’s AI capability could exacerbate growing tensions between the U.S. and China <span class="citation">(Zwetsloot, Toner, and Ding <a href="#ref-zwetsloot2018">2018</a>)</span>. As such, future research should explore how factual – non-exaggerated – information about American and Chinese AI capabilities influences public opinions.</p>
<div class="figure"><span id="fig:uschina"></span>
<img src="ai_public_opinion_us_2018_report-190107_web_files/figure-html/uschina-1.png" alt="Comparing Americans' perceptions of U.S. and China's AI research and development quality" width="2100" />
<p class="caption">
Figure 4.1: Comparing Americans’ perceptions of U.S. and China’s AI research and development quality
</p>
</div>
</div>
<div id="subsecexperimentchina" class="section level2">
<h2><span class="header-section-number">4.2</span> Communicating the dangers of a U.S.-China arms race requires explaining policy trade-offs</h2>
<!--
BZ: I think this subsection is about 85% there.
Can you make sure that you approve of my framing? My take on the results is that people do not understand there is a tradeoff between racing ahead and cooperation. People think that the U.S. could do both. The risk of war information treatment works because it makes people understand the trade off between the two policies.
-->
<p>In this survey experiment, respondents were randomly assigned to consider different arguments about a U.S.-China arms race. (Details of the survey experiment are found in <a href="apptopline.html#armsraceexp">Appendix B</a>.) All respondents were given the following prompt:</p>
<blockquote>
<p>Leading analysts believe that an AI arms race is beginning, in which the U.S. and China are investing billions of dollars to develop powerful AI systems for surveillance, autonomous weapons, cyber operations, propaganda, and command and control systems.</p>
</blockquote>
<p>Those in the treatment condition were told they would read a short news article. The three treatments were:</p>
<ol style="list-style-type: decimal">
<li><p><strong>Pro-nationalist treatment</strong>: The U.S. should invest heavily in AI to stay ahead of China; quote from a senior National Security Council official</p></li>
<li><p><strong>Risks of arms race treatment</strong>: The U.S.-China arms race could increase the risk of a catastrophic war; quote from Elon Musk</p></li>
<li><p><strong>One common humanity treatment</strong>: The U.S.-China arms race could increase the risk of a catastrophic war; quote from Stephen Hawking about using AI for the good of all people rather than destroying civilization</p></li>
</ol>
<p>Respondents were asked to consider two statements and indicate whether they agree or disagree with them:</p>
<ul>
<li><p>The U.S. should invest more in AI military capabilities to make sure it doesn’t fall behind China’s, even if doing so may exacerbate the AI arms race.</p></li>
<li><p>The U.S. should work hard to cooperate with China to avoid the dangers of an AI arms race, even if doing so requires giving up some of the U.S.’s advantages. Cooperation could include collaborations between American and Chinese AI research labs, or the U.S. and China creating and committing to common safety standards for AI.</p></li>
</ul>
<p>Americans, in general, weakly agree that the U.S. should invest more in AI military capabilities <em>and</em> cooperate with China to avoid the dangers of an AI arms race, as seen in Figure <a href="ai-policy-and-u-s-china-relations.html#fig:armsrace">4.2</a>. Many respondents do not think that the two policies are mutually exclusive. The correlation between responses to the two statements, unconditional on treatment assignment, is only -0.05. In fact, 29% of those who agree that the U.S. and China should cooperate also agree that the U.S. should invest more in AI military capabilities. (See Figure <a href="addresults.html#fig:armsracecorrfig">C.2</a> for the conditional percentages.)</p>
<p>Respondents assigned to read about the risks of an arms race (Treatment 2) indicate significantly higher agreement with the pro-cooperation statement (Statement 2) than the investing in AI military capabilities statement (Statement 1), according to Figure <a href="ai-policy-and-u-s-china-relations.html#fig:armsracediff">4.4</a>. Those assigned to Treatment 2 are more likely to view the two statements as mutually exclusive. In contrast, respondents assigned to the other conditions indicate similar levels of agreement with both statements.</p>
<p>After estimating the treatment effects, we find that the experimental messages do little to change the respondents’ preferences. Treatment 2 is the one exception. Treatment 2 decreases respondents’ agreement with the statement that the U.S. should invest more in AI military capabilities by 27%, as seen in Figure <a href="ai-policy-and-u-s-china-relations.html#fig:armsraceregression">4.3</a>. Future research could focus on testing more effective messages, such as op-eds <span class="citation">(Coppock et al. <a href="#ref-coppock2018long">2018</a>)</span> or videos <span class="citation">(Paluck et al. <a href="#ref-paluck2015does">2015</a>)</span>, which explains that U.S.’s investment in AI for military use will decrease the likelihood of cooperation with China.</p>
<div class="figure"><span id="fig:armsrace"></span>
<img src="ai_public_opinion_us_2018_report-190107_web_files/figure-html/armsrace-1.png" alt="Responses from U.S.-China arms race survey experiment" width="2100" />
<p class="caption">
Figure 4.2: Responses from U.S.-China arms race survey experiment
</p>
</div>
<div class="figure"><span id="fig:armsraceregression"></span>
<img src="ai_public_opinion_us_2018_report-190107_web_files/figure-html/armsraceregression-1.png" alt="Effect estimates from U.S.-China arms race survey experiment" width="2100" />
<p class="caption">
Figure 4.3: Effect estimates from U.S.-China arms race survey experiment
</p>
</div>
<div class="figure"><span id="fig:armsracediff"></span>
<img src="ai_public_opinion_us_2018_report-190107_web_files/figure-html/armsracediff-1.png" alt="Difference in response to the two statements by experimental group" width="2100" />
<p class="caption">
Figure 4.4: Difference in response to the two statements by experimental group
</p>
</div>
</div>
<div id="americans-see-the-potential-for-u.s.-china-cooperation-on-some-ai-governance-challenges" class="section level2">
<h2><span class="header-section-number">4.3</span> Americans see the potential for U.S.-China cooperation on some AI governance challenges</h2>
<!--
BZ: I think this subsection is pretty solid; about 90% there. I could pass it along to Jeff to see if there are other things I could cite.
-->
<p>We examined issue areas where Americans perceive likely U.S.-China cooperation. Each respondent was randomly assigned to consider three out of five AI governance challenges. For each challenge, the respondent was asked, “For the following issues, how likely is it that the U.S. and China can cooperate?”. (See <a href="apptopline.html#uschinacoop">Appendix B</a> for the question text.)</p>
<p>On each of these AI governance issues, Americans see some potential for U.S.-China cooperation, according to Figure <a href="ai-policy-and-u-s-china-relations.html#fig:coopchina">4.5</a>. U.S.-China cooperation on value alignment is perceived to be the most likely (48% mean likelihood). Cooperation to prevent AI-assisted surveillance that violates privacy and civil liberties is seen to be the least likely (40% mean likelihood) – an unsurprising result since the U.S. and China take different stances on human rights.</p>
<p>Despite current tensions between Washington and Beijing, the Chinese government, as well as Chinese companies and academics, have signaled their willingness to cooperate on some governance issues. These include banning the use of lethal autonomous weapons <span class="citation">(Kania <a href="#ref-kania2018">2018</a>)</span>, building safe AI that is aligned with human values <span class="citation">(China Institute for Science and Technology Policy at Tsinghua University <a href="#ref-chinaai2018">2018</a>)</span>, and collaborating on research <span class="citation">(News <a href="#ref-borderlessresearch">2018</a>)</span>. Most recently, the major tech company Baidu became the first Chinese member of the Partnership on AI, a U.S.-based multi-stakeholder organization committed to understanding and discussing AI’s impacts <span class="citation">(Cadell <a href="#ref-cadell2018">2018</a>)</span>.</p>
<p>In the future, we plan to survey Chinese respondents to understand how they view U.S.-China cooperation on AI and what governance issues they think the two countries could collaborate on.</p>
<div class="figure"><span id="fig:coopchina"></span>
<img src="ai_public_opinion_us_2018_report-190107_web_files/figure-html/coopchina-1.png" alt="Issue areas for possible U.S.-China cooperation" width="2100" />
<p class="caption">
Figure 4.5: Issue areas for possible U.S.-China cooperation
</p>
</div>
</div>
</div>
<h3>References</h3>
<div id="refs" class="references">
<div id="ref-cadell2018">
<p>Cadell, Cate. 2018. “Search engine Baidu becomes first China firm to join U.S. AI ethics group.” Reuters. <a href="https://perma.cc/DC63-9CFT">https://perma.cc/DC63-9CFT</a>.</p>
</div>
<div id="ref-chinaai2018">
<p>China Institute for Science and Technology Policy at Tsinghua University. 2018. “China AI Development Report 2018.” China Institute for Science; Technology Policy at Tsinghua University.</p>
</div>
<div id="ref-coppock2018long">
<p>Coppock, Alexander, Emily Ekins, David Kirby, and others. 2018. “The Long-Lasting Effects of Newspaper Op-Eds on Public Opinion.” <em>Quarterly Journal of Political Science</em> 13 (1): 59–87.</p>
</div>
<div id="ref-ding2018">
<p>Ding, Jeffrey. 2018. “Deciphering China’s AI Dream.” Technical report. Governance of AI Program, Future of Humanity Institute, University of Oxford. <a href="https://perma.cc/9BED-QGTU">https://perma.cc/9BED-QGTU</a>.</p>
</div>
<div id="ref-funk2015">
<p>Funk, Cary, and Lee Rainie. 2015. “Public and Scientists’ Views on Science and Society.” Survey report. Pew Research Center. <a href="https://perma.cc/9XSJ-8AJA">https://perma.cc/9XSJ-8AJA</a>.</p>
</div>
<div id="ref-goldfarb2018ai">
<p>Goldfarb, Avi, and Daniel Trefler. 2018. “AI and International Trade.” In <em>The Economics of Artificial Intelligence: An Agenda</em>. University of Chicago Press. <a href="https://www.nber.org/chapters/c14012.pdf">https://www.nber.org/chapters/c14012.pdf</a>.</p>
</div>
<div id="ref-chinausreport2017">
<p>IT Juzi and Tencent Institute. 2017. “2017 China-U.s. AI Venture Capital State and Trends Research Report.” Technical report. IT Juzi; Tencent Institute. <a href="https://perma.cc/AMN6-DZUV">https://perma.cc/AMN6-DZUV</a>.</p>
</div>
<div id="ref-kai2018ai">
<p>Kai-Fu, Lee. 2018. <em>AI Superpowers: China, Silicon Valley and the New World Order</em>. New York: Houghton Mifflin Harcourt.</p>
</div>
<div id="ref-kania2018">
<p>Kania, Elsa. 2018. “China’s Strategic Ambiguity and Shifting Approach to Lethal Autonomous Weapons Systems.” Lawfare. <a href="https://perma.cc/K68L-YF7Q">https://perma.cc/K68L-YF7Q</a>.</p>
</div>
<div id="ref-borderlessresearch">
<p>News, Bloomberg. 2018. “China Calls for Borderless Research to Promote AI Development.” Bloomberg News. <a href="https://www.bloomberg.com/news/articles/2018-09-17/china-calls-for-borderless-research-to-promote-ai-development">https://www.bloomberg.com/news/articles/2018-09-17/china-calls-for-borderless-research-to-promote-ai-development</a>.</p>
</div>
<div id="ref-paluck2015does">
<p>Paluck, Elizabeth Levy, Paul Lagunes, Donald P Green, Lynn Vavreck, Limor Peer, and Robin Gomila. 2015. “Does Product Placement Change Television Viewers’ Social Behavior?” <em>PloS One</em> 10 (9): e0138610.</p>
</div>
<div id="ref-west2018worries">
<p>West, Darrell M. 2018b. “Brookings survey finds worries over AI impact on jobs and personal privacy, concern U.S. will fall behind China.” Survey report. Brookings Institution. <a href="https://perma.cc/HY38-3GCC">https://perma.cc/HY38-3GCC</a>.</p>
</div>
<div id="ref-zwetsloot2018">
<p>Zwetsloot, Remco, Helen Toner, and Jeffrey Ding. 2018. “Beyond the Ai Arms Race.” <em>Foreign Affairs</em> November. <a href="https://perma.cc/CN76-6LG8">https://perma.cc/CN76-6LG8</a>.</p>
</div>
</div>
</section>
</div>
</div>
</div>
<a href="public-opinion-on-ai-governance.html" class="navigation navigation-prev " aria-label="Previous page"><i class="fa fa-angle-left"></i></a>
<a href="trend-across-time-attitudes-toward-workplace-automation.html" class="navigation navigation-next " aria-label="Next page"><i class="fa fa-angle-right"></i></a>
</div>
</div>
<script src="libs/gitbook-2.6.7/js/app.min.js"></script>
<script src="libs/gitbook-2.6.7/js/lunr.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-search.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-sharing.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-fontsettings.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-bookdown.js"></script>
<script src="libs/gitbook-2.6.7/js/jquery.highlight.js"></script>
<script>
gitbook.require(["gitbook"], function(gitbook) {
gitbook.start({
"sharing": {
"github": false,
"facebook": true,
"twitter": true,
"google": false,
"linkedin": false,
"weibo": false,
"instapaper": false,
"vk": false,
"all": ["facebook", "google", "twitter", "linkedin", "weibo", "instapaper"]
},
"fontsettings": {
"theme": "white",
"family": "sans",
"size": 2
},
"edit": null,
"history": {
"link": null,
"text": null
},
"download": [["https://governanceai.github.io/US-Public-Opinion-Report-Jan-2019/us_public_opinion_report_jan_2019.pdf", "PDF"], ["https://ssrn.com/abstract=3312874", "SSRN"], ["https://doi.org/10.7910/DVN/SGFRYA", "Replication Data"]],
"toc": {
"collapse": "section",
"scroll_highlight": true
},
"toolbar": {
"position": "fixed"
},
"search": false
});
});
</script>
<!-- dynamically load mathjax for compatibility with self-contained -->
<script>
(function () {
var script = document.createElement("script");
script.type = "text/javascript";
var src = "true";
if (src === "" || src === "true") src = "https://mathjax.rstudio.com/latest/MathJax.js?config=TeX-MML-AM_CHTML";
if (location.protocol !== "file:" && /^https?:/.test(src))
src = src.replace(/^https?:/, '');
script.src = src;
document.getElementsByTagName("head")[0].appendChild(script);
})();
</script>
</body>
</html>