PLOS ONE

RESEARCH ARTICLE

Exploring the artificial intelligence “Trust

paradox”: Evidence from a survey experiment
in the United States
Sarah Kreps1, Julie George ID1,2*, Paul Lushenko1, Adi Rao1

1 Cornell University Tech Policy Institute, Menlo Park, CA, United States of America, 2 Stanford Center for
International Security and Cooperation, Stanford, CA, United States of America

* [email protected]

a1111111111
a1111111111 Abstract
a1111111111
a1111111111 Advances in Artificial Intelligence (AI) are poised to transform society, national defense, and
a1111111111
the economy by increasing efficiency, precision, and safety. Yet, widespread adoption within
society depends on public trust and willingness to use AI-enabled technologies. In this
study, we propose the possibility of an AI “trust paradox,” in which individuals’ willingness to
use AI-enabled technologies exceeds their level of trust in these capabilities. We conduct a
OPEN ACCESS
two-part study to explore the trust paradox. First, we conduct a conjoint analysis, varying dif-
Citation: Kreps S, George J, Lushenko P, Rao A ferent attributes of AI-enabled technologies in different domains—including armed drones,
(2023) Exploring the artificial intelligence “Trust
general surgery, police surveillance, self-driving cars, and social media content moderation
paradox”: Evidence from a survey experiment in
the United States. PLoS ONE 18(7): e0288109. —to evaluate whether and under what conditions a trust paradox may exist. Second, we
https://doi.org/10.1371/journal.pone.0288109 use causal mediation analysis in the context of a second survey experiment to help explain
Editor: Hans H. Tung, National Taiwan University, why individuals use AI-enabled technologies that they do not trust. We find strong support
TAIWAN for the trust paradox, particularly in the area of AI-enabled police surveillance, where the lev-
Received: January 20, 2023 els of support for its use are both higher than other domains but also significantly exceed
trust. We unpack these findings to show that several underlying beliefs help account for pub-
Accepted: June 20, 2023
lic attitudes of support, including the fear of missing out, optimism that future versions of the
Published: July 18, 2023
technology will be more trustworthy, a belief that the benefits of AI-enabled technologies
Copyright: © 2023 Kreps et al. This is an open outweigh the risks, and calculation that AI-enabled technologies yield efficiency gains. Our
access article distributed under the terms of the
findings have important implications for the integration of AI-enabled technologies in multiple
Creative Commons Attribution License, which
permits unrestricted use, distribution, and settings.
reproduction in any medium, provided the original
author and source are credited.

Data Availability Statement: We have uploaded

the minimal anonymized data set necessary to
replicate our study findings to a stable, public
repository. This is the link where one can access Introduction
that information: https://doi.org/10.7910/DVN/ In August 2022, The New York Times observed “we’re in a golden age of progress in artificial
EOYDJR.
intelligence (AI). It’s time to start taking its potential and risks seriously” [1]. Amid these rapid
Funding: The author(s) received no specific developments in AI, discussion of human agency is often absent. This oversight is puzzling
funding for this work. given that individuals will ultimately be responsible for whether AI-enabled technologies dif-
Competing interests: The authors have declared fuse widely across society or confront resistance. Regulators will have some role to play, and
that no competing interests exist. researchers have studied local officials’ reactions to AI-enabled technologies precisely because

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 1 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

they will make important decisions about if, when, and how to use these capabilities [2]. Not-
withstanding these insights, we contend that public attitudes are crucial to societal adoption of
AI-enabled technologies. If the public is hesitant, demonstrating reluctance to adopt these
technologies, then it will pressure policymakers to impose restrictions on research and devel-
opment [3].
In this study, we research public attitudes both toward AI-enabled technologies across vari-
ous domains and the basis for those attitudes. Existing studies about public attitudes toward
emerging technologies have tended to focus on levels of trust, which is defined as the “willing-
ness to make oneself vulnerable” to a capability “based on a judgment of similarity of inten-
tions or values” [4]. Research on nanotechnology focuses on the public’s degree of trust in the
capability to minimize risks to humans [5, 6], as does research on genetically modified organ-
isms [7] and online shopping [8]. Studies on AI-enabled technologies are no different, with
researchers preferring to use trust as the dependent variable [9]. This focus on trust may not
be misguided. In the context of nuclear energy, for instance, researchers have studied public
trust, in part because previous work shows a strong relationship between low levels of trust
and low public support for nuclear power [10]. Researchers have also shown that public trust
for autonomous vehicles correlates with greater levels of acceptance and adoption [11], which
is similar for public attitudes regarding mobile banking [12].
We draw on this scholarship to suggest and test the possibility of a “trust paradox,” which
we define as the public’s puzzling willingness to support or use AI-enabled technologies that
they do not trust. Such a dynamic is well documented in the social media space [13–15], where
the public heavily uses social media despite expressing concerns about data privacy, content
moderation, and misinformation. Why would the public support the use of AI-enabled tech-
nologies it does not trust? In addressing this question, we advance five hypotheses that help
explain the puzzle of why individuals support the use of AI-enabled technologies despite hav-
ing lower levels of trust: the “fear of missing out” (FOMO); a cost-benefit analysis wherein
individuals see risk but are persuaded by the potential benefits; assessments about the absence
of efficient alternatives; optimism about the future development of more trustworthy technol-
ogy; and, transparency about the nature of technology.
First, one potential explanation for varying levels of support and trust of AI-enabled tech-
nologies centers on FOMO. The literature on consumer psychology points to FOMO as a pow-
erful factor that can influence people to embrace an experience or purchase a product, even if
they believe that doing so can be self-defeating. This FOMO mechanism has been cited as the
reason why individuals often overuse smartphones, sleep too little, or abuse drugs, despite rec-
ognizing that such behavior contradicts privately-held beliefs or values [16]. We apply this
insight, which is related to the implications of other social-psychological considerations such
as status [17] and reputation [18] on personal behavior [19], to AI. We argue that FOMO may
explain why individuals support the use of AI-enabled technologies that they do not trust. Peo-
ple may believe that if they do not use a certain AI-enabled technology, others will, resulting in
feelings of anxiety or loss because they are somehow missing out on seemingly popular
activities.
Second, the public’s trust paradox for AI-enabled technologies may also result from a calcu-
lation that while these technologies introduce risk, they also provide more benefits overall.
Indeed, people may rely on AI-enabled technologies not so much because they trust these
capabilities but because they perceive that the anticipated benefits [20] of adopting AI-enabled
technologies will exceed the expected costs [21] of doing so. This amounts to an expected util-
ity calculation about the benefits and costs of behavior characteristic of the risk management
[3] of new technologies, in which the risks are assessed relative to overall benefits and adopted
if the latter outweighs the former. In the context of AI-enabled technologies, this calculation

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 2 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

might also help explain individuals’ support of new capabilities that they do not trust, reflect-
ing a belief that they stand to enjoy benefits that offset their perception of costs in the face of
distrust. While individuals may not trust an AI-enabled technology, they may still understand
it as conferring improvements over a human-driven alternative.
Third, individuals might support the use of technologies at rates that exceed trust if they see
few economically viable alternatives. For example, AI-enabled technologies can be used as a
substitute for human labor across a range of tasks, especially through automation [22]. Bryn-
jolfsson and McAfee [23] have traced the way that automation has increasingly added efficien-
cies to the economy, performing medical diagnoses, replacing humans on assembly lines [22,
24–27], and carrying loads too heavy for humans. In other examples, accounting, sales, and
trading-related tasks are being replaced by AI [28]. These developments create thorny issues of
trust and privacy, but also might seem inevitable, as emerging technologies in the past have
also remade societies and economies [26, 27]. Individuals, then, might share these concerns
but nonetheless acknowledge that AI-enabled technology can perform particular jobs better
than humans, and potentially, replace these human jobs.
Fourth, we posit that technological optimism may be another explanation. In this case, indi-
viduals may believe that even if they face risks in the present, future iterations of an AI-enabled
technology will improve in ways that minimize such potential harms, which is consistent with
technological improvements in the past [29]. A previous study of individual technology use
showed that whereas many individuals believe that digital capitalism currently disadvantages
large swaths of society as well as erodes trust in democratic institutions, they also think future
technology will provide solutions to these challenges through the protection of speech and
empowerment of citizens [30]. Based on these findings, individuals might not trust AI-enabled
technologies now, but have confidence that these capabilities will improve over time. Such
technological optimism may encourage them to adopt AI-enabled technologies in the short-
term given the promise of longer-term improvements.
Finally, individuals may not trust AI-enabled technologies but support their use if there is
transparency or explanation for how these technologies are used. One of the reasons people
have distrusted AI is because the enabling algorithm is perceived as a “black box.” The lack of
explanation for coding decisions as well as datasets to train the enabling algorithms creates the
potential for bias in AI [31]. This helps explain why Twitter’s CEO, Elon Musk, recently stated
“transparency is key to trust” [32], echoing Putnam’s findings that trust is integral to demo-
cratic society [33]. Thus, Musk promised to better communicate the social media platform’s
data management protocols because of concerns that Twitter “interfered” in the 2020 U.S.
presidential election [34]. In the context of AI-enabled technologies, then, the expectation of
transparency may account for the public’s willingness to use these systems that they otherwise
distrust [35].

Materials and methods

To study the potential differences between public attitudes in terms of support and trust, both
within and across different applications of AI, we constructed a two-part empirical study con-
ducted on a representative sample of 1,008 U.S. citizens. Our project received an exempt status
from the Cornell University’s Institutional Review Board for Human Subject Research (IRB
Protocol #2004009569), and we obtained consent from respondents electronically. The form
of consent obtained was electronic. No minors were included in this study. Subjects were
properly instructed and have indicated that they consent to participate in our study by elec-
tronically approving the appropriate informed consent. First, although observational evidence
points to a trust paradox in some contexts, we seek to establish whether this is a generalized

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 3 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

phenomenon across AI-enabled technologies by fielding a conjoint survey that assesses per-
ceptions of support and trust in these technologies.
Second, building on the five mechanisms outlined above, we explore the reasons why indi-
viduals exhibit differences in support and trust using causal mediation analysis. We adminis-
tered the study from October 7–21, 2022, via Lucid. Lucid uses a quota sampling protocol to
produce samples that correspond to U.S. demographics in terms of age, gender, race, and loca-
tion. Existing research shows that this sampling protocol produces experimental effects that
largely mirror those found with probability-based sampling [36]. We present summary statis-
tics in the Supplementary Information.
As Fig 1 shows, respondents first participated in a conjoint experiment to investigate the
potential gap between support and trust in emerging technology. Marketing surveys

Fig 1. Survey flowchart of conjoint and causal mediation analyses via lucid (October 7–21, 2022).
https://doi.org/10.1371/journal.pone.0288109.g001

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 4 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

commonly use conjoint surveys [37] because they enable researchers to vary a number of attri-
butes and assess how levels of those attributes affect individual choice. Research also finds that
conjoint surveys help reduce social desirability bias among respondents, in which they are
encouraged to answer in a certain way or do so because they feel obligated [38]. In adopting
this approach, we ensure to fulfill a key assumption for the orthogonality of attributes in expec-
tation, which enables us to differentiate complex treatment effects into their constituent parts
[39, 40]. In our case, we are able to assess multiple factors that could plausibly relate to the
anticipated effects of AI attributes, either discretely or when interacted with each other, on
respondents’ preference formation.
Initially, we informed respondents that they would be presented with a series of hypotheti-
cal applications of AI-enabled technologies. We varied four different attributes, outlined in
Table 1, along with their respective levels [41]. Research suggests that hypothetical but realistic
combinations of attributes can help enhance the external validity of survey results by calibrat-
ing contextual detail with experimental control [42, 43].
In the first part of our experiment, the conjoint survey, we consider different domains in
which AI operates. Research on attitudes toward AI typically focuses on one domain, such as
autonomous vehicles [44]. We advance the literature on individual preferences for AI-enabled
technologies by considering how respondents may have different perceptions of risk across
domains, which allows us to understand—theoretically and empirically—how receptiveness to
AI varies depending on context. We select domains where AI is already making inroads,
including autonomous vehicles, armed drones, general surgery, social media content modera-
tion, and police surveillance. In doing so, we draw on categories identified in the One Hundred
Year Study of Artificial Intelligence published by Stanford University, which studies how AI
will affect society in the future [45].
Within the conjoint survey we also evaluate the degree of autonomy. Drawing on the auton-
omy literature [46–48], we locate a range of autonomy that is bookended by extreme variations
and includes a hybrid form of automation as well. At one extreme, we present respondents a
form of manual autonomy that is characterized by human control with no machine assistance
or involvement. At the other extreme, we present respondents with a form of full autonomy
where the machine has full responsibility and control over decision-making in these settings.
Splitting the difference between these two extreme variations in autonomy is mixed-initiative

Table 1. Considered attributes and attribute levels for artificial intelligence applications.
AI Attributes Levels
Domain Cars
Armed drones
General surgery
Social media content moderation
Police surveillance
Degree of Fully autonomous (no human in the loop)
autonomy Mixed-initiative (on-/off switch between human and machine)
Manual (no autonomy)
Precision Maximum precision—a model that produces up to no false positives, and is correct 99% of the
time.
Substantial precision—a model that produces 10% false positives, or in other words, is correct
90% of the time.
Moderate precision—a model that produces 15% false positives, or in other words, is correct
85% of the time.
Regulation Public government agencies
Private industry
Individual or community users
https://doi.org/10.1371/journal.pone.0288109.t001

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 5 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

autonomy. Respondents presented with this level of autonomy are informed that the applica-
tion can toggle between human and machine control in these settings.
We then vary the degree of algorithmic precision. Model performance is often a function of
convergence between the algorithm and truth. In a context of natural language models, for
example, algorithmic detection of AI-generated text is measured based on the percentage of
text examples that the algorithm correctly identified as AI versus human generated [47]. Social
media platforms such as Facebook advertise that they remove 99.8% of terrorism-related con-
tent before users flag it, and 97% for hate speech that violates community standards [49]. Algo-
rithms trained to detect cancer are measured in terms of accuracy of diagnoses, with some
neural networks reaching 99.8% [50]. In our study, we use language concerning the precision
of algorithms, which refers to the percentage of the results which are relevant. Designing an
algorithm for precision yields confident diagnoses. In the context of oncology, this means that
an algorithm correctly diagnoses someone with cancer while avoiding Type II errors or false
positives. A model that has the precision of 0.95 when predicting a diagnosis, then, would be
correct 95% of the time. How individuals view these precision rates is likely to be a function of
the domain in which algorithms operate, interacting with variables such as the controllability
to affect what they deem to be “sufficiently high” levels of precision.
We additionally consider the locus of regulation. One approach is for government agencies
to regulate the way that businesses use algorithms by enforcing legislation such as the Credit
Reporting Act. A second option is for a private firm that developed the software or platform to
regulate it, similar to the approach adopted by Facebook and Twitter. A third approach is pub-
lic-private collaboration where public agencies direct private firms to alter their behavior. The
Centers for Disease Control often request that Twitter flag particular content relating to
COVID-19 vaccines as misinformation [51]. Finally, individuals, users, or communities engag-
ing on a particular social media platform may self-regulate, which is consistent with Reddit’s
approach to content management. Reddit communities adopt this approach in terms of crea-
tors developing guidelines for user behavior and communities regulating on subreddits [52].
Taken together, the conjoint task randomly varied four attributes and levels therein, yield-
ing 135 unique scenarios for an AI-enabled technology, which a recent study shows is well
within the tolerance level for quality responses [53]. Given our within-subject survey design,
we presented respondents with four randomly assigned choice sets resulting in over 5,000
observations. After reading each scenario, we then asked respondents two questions that
define our key dependent variables. First, we ask whether respondents support the use of AI in
these settings, using a 5-point Likert scale to gauge their attitudes (1 corresponds to “strongly
disagree” and 5 to “strongly agree”). To gain leverage over a possible trust paradox where
respondents do not trust an AI-enabled technology but nevertheless support its adoption, we
ask subjects if they trust the use of AI for these purposes as well. We study trust as one of our
outcome variables rather than a willingness to use AI-enabled technologies for several reasons.
Measuring trust is consistent with previous research for emerging technologies and more
importantly, is a direct test of public attitudes. Measuring respondents’ willingness to use AI-
enabled technologies, on the other hand, may increase the potential for confirmation bias
regarding these capabilities.
To analyze our data, we first evaluated the potential for a trust paradox, initially calculating
marginal means for the degree of support and trust for each AI-attribute level in a manner
similar to other conjoint studies in political science [54]. Doing so allows us to identify the
“trust paradox,” which we do by determining the statistical difference between public support
and trust for each AI-attribute level. We then calculate the average marginal component effect
(ACME) per attribute level, which is generated by our conjoint design, in a regression frame-
work. The ACME represents the mean difference in a respondents’ level of support and trust

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 6 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

for an AI-enabled technology when comparing two attribute values averaged across all combi-
nations of other AI-enabled technology attribute values. Based on the randomization of our
attribute-levels, we assume that these ACMEs are nonparametric, meaning they are not the
result of an underlying distribution of data [55–57]. We also include several control variables
in our regression framework based on existing research that suggests these may have poten-
tially important mediating effects on public attitudes for AI-enabled technologies. Specifically,
we are interested in how differences in age, gender, race, and political ideology may shape
respondents’ support and trust for AI-enabled technologies [58, 59].
Beyond establishing whether a trust paradox exists, we then presented respondents with
another survey experiment to assess the potential microfoundations, or underlying values and
beliefs developed in the introduction, that shape degrees of support and trust in AI-enabled
technology. This consisted of a 3x2 factorial and between subject survey including six treat-
ments and a control group. Those randomly assigned to the control group only learned that
“In recent years, advancements in Artificial-Intelligence (AI) have led to the emergence of new
technologies.” From the five domains we studied in our conjoint survey, we selected three that
scholars broadly recognize as the most “hotly debated” [45]: armed drones, social-media con-
tent moderation, and driverless vehicles. These three domains also capture the use of AI-
enabled technologies in various settings, such as in conflict (armed drones), across society
(driverless cars), and in the online space (social media content moderation), which further
allows us to investigate potential differences in support and trust.
For each technology, the experimental groups also varied two intended purposes, whether
to enhance human judgment or provide a substitute for human judgment, which draws on
debates about “complementing” versus “replacing” human systems. One argument suggests
that AI-enabled technologies have different cognitive qualities than biological systems—
including a set of mental models about ethics and empathy—and should be viewed less as a
replacement for human intelligence but rather as a partner to humans [60].
After reading their vignettes, we asked respondents to gauge their support, trust, and
understanding for AI-enabled technologies with varying purposes. Previous research on medi-
ators in terms of AI-enabled technologies has focused largely on the implications of affect,
such as anger, on degrees of public support [61]. Rather, we cast a wider net of potential medi-
ators, building on arguments about trust and support in the literature on emerging technology
discussed in the introduction. Specifically, we draw on the five mechanisms outlined in the
introduction to develop corresponding statements and ask individuals their degree of agree-
ment or disagreement with each of these statements, using a 5-point Likert scale to capture
their feedback (1 corresponds to “strongly disagree” and 5 to “strongly agree). For the FOMO
mechanism, for example, we asked subjects to respond to the following statement: “Others are
using this AI-enabled technology, so not using it means the prospect of missing out” [16]. We
repeated this approach for the other potential mediators, as outlined above in Fig 1.
Based on the responses to these questions, we then carried out causal mediation analysis.
This method shows the complete causal chain for the effect of an independent variable on a
mediator and the effect of a mediator on the dependent variable [62, 63]. Though we ensure to
fulfill key assumptions to operationalize this method, namely randomizing the order of survey
questions for respondents across all groups [64], causal mediation analysis is sometimes criti-
cized for failing to account for confounding variables, even in an experimental setting that
researchers usually champion for resolving stochastic error. Miles argues that confounders
“cannot be eliminated even in a well-controlled randomized experiment” [65], causing Simon-
sohn to contend that “we over-estimate the mediators” [66]. In recognition of these valid con-
cerns, we draw on previous studies that attempt to adjudicate public attitudes for AI-enabled
technologies. These studies adopt what Imai et al. refer to as the “sequential ignorability

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 7 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

assumption,” whereby both possible pretreatment confounders and treatment assignment are
assumed to be statistically independent from the potential outcomes and mediators [67]. Addi-
tionally, we opt not to inductively derive possible mediators from respondents’ answers to
open-ended questions, as other researchers have done [68], given the possibility of bias [69].

Results
We report the results of our conjoint analysis assessing the trust paradox and then the experi-
mental treatments evaluating the underlying mechanisms of public attitudes. Fig 2 presents
the marginal means for each attribute and level for public attitudes defined in terms of support
and trust. Our analysis of the data points to statistically significant differences between the
trust in an AI-enabled technology and the support of its use across certain domains, and at dif-
ferent autonomy and precision levels. Below, we discuss the results of our conjoint analysis
and then model the data within a regression framework.

Fig 2. Support and trust marginal means by attribute levels. Caption: AI-enabled technology attributes and support
or trust for the technology. The figure shows the marginal means for each attribute value for support or trust in the
hypothetical AI-enabled technology, as indicated by respondents after reading their assigned scenarios, on a 5-point
scale (rescaled to vary from 0 to 1). Estimates are based on Table 2. Error bars represent 95% CIs around each point
estimate.
https://doi.org/10.1371/journal.pone.0288109.g002

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 8 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

Fig 2 shows significant gaps between trust and support across attributes, validating the
notion of a trust paradox in which individuals’ support for use exceeds that of their trust.
Below, we report the marginal mean sizes and statistical significance of gaps between support
and trust per attribute. These differences reflect around a 5% difference in levels of support
and trust. In terms of domain, we find that the trust paradox is largest for police surveillance
(-0.059 points; p < 0.001), followed by drones (-0.054 points; p < 0.001), cars (-0.050 points;
p < 0.001), general surgery (-0.047 points; p < 0.001), and social media content moderation
(-0.045 points; p < 0.001). We use a second survey experiment in conjunction with causal
mediation analysis to unpack these results, which we discuss in the section following. Turning
to autonomy, we see a significant decrease in trust relative to support for full human agency
(-0.052; p < 0.001), mixed autonomy (-0.051 points; p <0.001), and full autonomy (-0.049
points; p < 0.001). While the trust paradox is equally consistent for precision across the three
—85%, 90%, and 95%—levels, we find the gap is highest for the middle tier (-0.055 points;
p < 0.001). Lastly, in terms of regulatory agent, we also find gaps between support and trust,
whether the agent is an individual or community users (-0.046 points; p < 0.001); private
industry regulation (-0.050 points; p < 0.001); and, for public government agencies, trust is
also lower than support (-0.057 points; p < 0.001).
Next, we generated OLS regressions based on whether individuals support and trust the AI-
enabled technology. We present our results in Table 2. Model 1 shows support for the use of
the technology as the dependent variable. Model 2 replicates these results while using trust as
the dependent variable. Given the structure of our data, that is, our use of categorical variables,
we produce these results by using a referent level within each attribute, which are identified in
the table notes below.
Our analysis suggests that full autonomy significantly reduces respondents’ support (β =
-0.132, p < 0.01) and trust (β = -0.117, p < 0.01) for an AI-enabled technology. We note, how-
ever, that there is no effect when comparing the fully human system referent level to mixed
autonomy. In terms of domain, we see that, relative to the driverless cars referent level, police
surveillance significantly increases respondents’ support (β = 0.155, p < 0.01) and trust (β =
0.117, p < 0.01). We do not detect an effect for any other domain. These results, however, are
moderated by the precision. We find that 99% accuracy favorably shapes public support (β =
0.394, p < 0.001) and trust (β = 0.374, p < 0.001). For 90% accuracy, our analysis also shows
that support (β = 0.108, p < 0.01) is significant yet trust (β = 0.074, p < 0.05) is not. Similarly,
we do not find that variation in the regulatory agent shapes public attitudes of support and
trust. While respondents might care that some regulations exist, they are otherwise agnostic to
how different patterns of regulation may affect AI-enabled technologies.
We also assess how different demographic variables may impact levels of support and trust,
including race, age, gender, and education. We do not find significant differences in support
or trust for AI-enabled technologies on the basis of race. That we do not detect stronger effects
is somewhat surprising given the research on the range of biases that AI can have for commu-
nities of color, whether in terms of algorithms that discriminate against black patients on
transplant lists [70], in policing [71], and marketing [72]. Black individuals have been found to
express more distrust for AI-based facial recognition technology, perhaps because of the nega-
tive association with biased algorithms [73]. Our findings may point to less distrust for algo-
rithms than human judgment when it comes to policing but we urge additional study to probe
these attitudes more fully.
We do see, however, an association for age. Being older is associated with less support (β =
-0.008, p < 0.001) and trust (β = -0.008, p < 0.001) in AI-enabled technologies. This is consis-
tent with public opinion polling [74] that suggests millennials use technology more than older
generations, with 93% of millennials (23–38 years old) owning smartphones compared to

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 9 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

Table 2. Attributes and public preferences on ai-enabled technologies.

No Controls With Controls
Support Trust Support Trust
(Intercept) 3.265*** 3.096*** 3.392*** 3.201***
(0.063) (0.064) (0.132) (0.133)
Armed drones 0.052 0.037 0.054 0.039
(0.055) (0.054) (0.054) (0.053)
General surgery 0.067 0.079 0.080 0.093+
(0.053) (0.054) (0.051) (0.052)
Police surveillance 0.150** 0.113* 0.155** 0.117*
(0.055) (0.054) (0.054) (0.053)
Social media content moderation -0.002 0.019 -0.011 0.012
(0.055) (0.054) (0.053) (0.052)
Full autonomy and no human oversight -0.128** -0.114** -0.132** -0.117**
(0.043) (0.042) (0.042) (0.042)
Mixed autonomy (human-in-the-loop) 0.021 0.027 0.024 0.031
(0.041) (0.041) (0.040) (0.040)
Maximum precision (correct 99% of the time with 1% false positives) 0.386*** 0.366*** 0.394*** 0.374***
(0.041) (0.042) (0.040) (0.041)
Substantial precision (correct 90% of the time with 10% false positives) 0.122** 0.086* 0.108** 0.074+
(0.040) (0.040) (0.039) (0.039)
Private industry 0.016 0.002 0.015 0.001
(0.039) (0.040) (0.038) (0.039)
Public government agencies 0.038 -0.004 0.039 -0.004
(0.041) (0.041) (0.041) (0.041)
Male 0.188** 0.213***
(0.057) (0.057)
Conservatism -0.041* -0.037*
(0.018) (0.019)
Income 0.040+ 0.027
(0.022) (0.022)
Education 0.069** 0.067**
(0.023) (0.023)
White -0.094 -0.074
(0.062) (0.065)
Age -0.008*** -0.008***
(0.002) (0.002)
Num.Obs. 5040 5040 5040 5040
R2 0.024 0.021 0.071 0.062
R2 Adj. 0.022 0.019 0.068 0.059
RMSE 1.16 1.16 1.13 1.14
Std.Errors by: id by: id by: id by: id

+ p < 0.1
* p < 0.05
** p < 0.01
*** p < 0.001
Caption: Conjoint average marginal component effects (AMCE) per attribute level. We use cars, human only autonomy, 85% precision, and community and individual
regulations as referents for domain, autonomy, precision, and regulator respectively. The dependent variable is a 5-point Likert scale.

https://doi.org/10.1371/journal.pone.0288109.t002

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 10 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

older people (74–91 years old). There may also be differences based on gender, with men more
likely than women to support (β = 0.188, p < 0.001) and trust (β = 0.213, p < 0.001) the use of
AI-enabled technologies. These findings reflect earlier research for a gender-bias in terms of
AI, with men more likely to support and trust emerging technologies than women [75, 76].
Education is also positively associated with both support (β = 0.069, p < 0.01) and trust (β =
0.067, p < 0.01); however, conservatism appears to be negatively associated with support (β =
-0.041, p < 0.05) and trust (β = 0.037, p < 0.05). Similar to our finding for race, we approach
any interpretation of non-randomized variables with caution—we do, however, present inter-
esting avenues for future research along these domains.

Examining the basis for support

The initial part of our study found a significant trust paradox, with individuals supporting the
use of technologies at substantially higher rates (e.g., the difference of β = 0.155 and β = 0.117
for support and trust, respectively, for the association with police surveillance) than their levels
of trust. This corroborates our theoretical hunch but does not necessarily provide clues as to
why the public would demonstrate high levels of support for technologies they do not neces-
sarily trust. To probe the basis of that paradox, we conducted a second part of the study, which
investigated the factors that mediate the relationship between the technology and public atti-
tudes of support or trust.
We also investigated the extent to which respondents understand the AI-enabled technolo-
gies, especially since AI can be difficult to visualize or comprehend. Indeed, “black box AI” has
been used to describe an AI system whose inputs and operations are not visible to users [77].
Including questions regarding the understanding of AI-enabled technologies in different
domains allows us to proxy for respondents’ self-assessed knowledge as well as determine how
this relates to levels of support and trust. These questions on understanding were asked after
the questions on support and trust. It is possible, then, that respondents may demonstrate
degrees of understanding that are at odds with their levels of support and trust, suggesting
important implications for the regulation of technologies.
This second experiment is not directly comparable to our first, meaning we do not directly
study the “trust paradox” identified above. Rather, we use this second experiment to gain
insights into the role that underlying beliefs may play in shaping overall attitudes for support
and trust for AI-enabled technologies, which we show can and often do deviate. In doing so,
we replicate an approach adopted by other scholars studying emerging technologies [78, 79].
Fig 3 shows the decline in support (Panel A) and trust (Panel B) for each domain of AI-enabled
technology and depending on whether the technology enhances or substitutes for human judg-
ment, relative to the control condition, which referenced generic advancements in AI leading
to new technologies. As the figure shows, the domain of social media reduced trust and sup-
port the most of any treatment condition in terms of enhance and substitute for human judg-
ment, and by nearly the same margin. Social media content moderation to substitute for
human judgment decreased trust and support by 18.73% (p < 0.001) and 18.48% (p < 0.001),
respectively, relative to the control group. The difference between levels of support for social
media content moderation that enhances rather than substitutes for human judgment is also
statistically significant (8.87%, p < 0.05), which is similar to changes in levels of trust (7.53%,
p < 0.05).
Drones and cars, on the other hand, decreased trust and support at lower levels relative to
the control and across both types of AI purposes—enhance human judgment and substitute
for human judgment—and at statistically-significant levels (p < 0.001). Whereas the public
shows less trust than support for cars in terms of AI-enabled technology that both enhances

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 11 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

Fig 3. Support (Panel A) and Trust (Panel B) for Treatment Conditions Relative to the Control. Caption: Values
represent changes in levels of support and trust for AI-enabled technologies by treatment conditions compared to the
control condition. Values are negative because levels of support and trust drop compared to the control or baseline
group.
https://doi.org/10.1371/journal.pone.0288109.g003

and substitutes for human judgment, the results are reversed for drones. The public shows
more trust for drones relative to the control group, in terms of AI that both enhances and sub-
stitutes for human judgment, than it does for support. At the same time, the public is more apt
to trust and support drones—relative to the control group—when AI is used to substitute for
human judgment rather than enhance human judgment. This may suggest that people have
limited knowledge of how drones work, which still require a “human-in-the-loop” to authorize
strikes; believe that drones incorporate more AI than they actually do; or, are more receptive
to fully autonomous drones than most scholars believe [80].
We also evaluated the degree to which individuals understand these technologies, recording
responses to the question “I have sufficient understanding of AI and how it works across
domains.” As Fig 4 below suggests, the relationship is nearly inverted relative to the trust and
support dependent variables in Fig 3 above. Compared to the baseline or control group, indi-
viduals seem to believe they understand social media content moderation less compared to the
control condition of generic AI-enabled technology than for AI-enabled cars, with cars regis-
tering more understanding than the control for both enhance (10.14%, p < 0.001) and substi-
tute (5.72%, p < 0.10) for human judgment. This evidence is consistent with observational
data showing that 94% of Americans are aware of efforts to develop autonomous or driverless
vehicles [81]. Indeed, approximately 92% of new vehicles have some driver assist function
(e.g., automated speed through adaptive cruise control) that is based on AI [82], which may
account for the increased understanding of AI-enhanced cars.
Social media content moderation showed the lowest levels of understanding for both
enhance (0.86%, p < 0.81) and substitution (4.76%, p < 0.14) of human judgment. This find-
ing is consistent with previous studies showing that algorithmic content moderation online
remains opaque and difficult to audit or understand [83]. Similar to our results above, we also

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 12 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

Fig 4. Change in understanding for treatment conditions relative to the control. Caption: Values represent changes
in levels of support and trust for AI-enabled technologies by treatment conditions compared to the control condition.
https://doi.org/10.1371/journal.pone.0288109.g004

observe that the public seems to believe they understand AI-enabled drones that substitute for
human judgment (6.13%, p < 0.05) more than other AI-enabled technologies, which may
reflect broader societal narratives on the future ubiquity of military robots on the battlefield
[84].
Modelling these results in a regression framework, both with and without the incorporation
of relevant controls to capture variation across members of the public, also reflects lower levels
of support and trust for social media content moderation compared to other AI-enabled tech-
nologies (see Supplemental Information for the regression output tables). These results are
consistent in both the basic and full regression models and for the use of social media content
moderation to enhance and substitute for human judgment. In the full model, for example,
social media content moderation to substitute for human judgment results in the lowest degree
of support (β = -0.66, p < 0.01) and trust (β = -0.64, p < 0.01) of any experimental treatment.
To help explain these results, paying special attention to social media content moderation
to substitute for human judgment (experimental group four), we then adopted causal media-
tion analysis. To calculate the proportion of the indirect treatment effect on the outcome
explained by our five mediators, we multiplied the effect of the treatment on the mediators by
the effect of the mediators on the outcome and then divided by the total treatment effect. This
approach replicates a method proposed [85] and adopted by other scholars [64, 86, 87].
As indicated in Fig 5 below, we found that those assigned to experimental group four were
less likely to say that they believe the appropriate guardrails may not be in place today but they
feel confident that those will be in place in the future (13.7%, p < 0.01). Similarly, those in
experimental group four were 12% (p < 0.02) less likely to report that AI was an effective sub-
stitute for humans doing the same task, 15.9% (p < 0.002) less likely to report that AI might
introduce risks but would nonetheless be more likely to generate overall benefits, and 13.9%
(p < 0.008) less likely to report a fear of missing out. We found that transparency did not

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 13 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

Fig 5. Factors mediating trust in ai-enabled technology. Caption: The top graph is for support, and the second graph
is for trust. 95% confidence intervals are in whiskers about each point estimate. Whiskers that cross the dashed line at
zero denote non-statistically significant results. Some values are negative because the effect of the mechanism acts in
the opposite direction of the overall treatment effect.
https://doi.org/10.1371/journal.pone.0288109.g005

exercise a statistically significant shaping effect on overall attitudes of support and trust, indi-
cating that the public emphasizes this consideration less than other factors when determining
their use of AI-enabled technologies. Together, then, four of the potential five mediators exer-
cise a negative indirect effect on the overall treatment effect, meaning respondents are not gen-
erally hopeful about the merits of AI-enabled technologies in terms of social media content
moderation that substitutes for human judgment for a variety of reasons.
Though useful, the Baron and Kenney [85] method has no significance test. To assess the
strength of the results, we ran a robustness check using the Sobel-Goodman mediation test. In
effect, these are T-Tests that show whether the indirect effect of the mediators on the overall
treatment effect is statistically different from zero. The Sobel-Goodman mediation tests show

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 14 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

significant mediating effects for FOMO (p = 0.01), calculation (p = 0.003), efficiency

(p = 0.02), and optimism (p = 0.007) in terms of support, and nearly identical results in terms
of trust. Consistent with our findings above, the mediating effect for transparency was not sta-
tistically significant. Therefore, we can be reasonably assured that four of the five mediators
exercise some effect on overall attitudes of support and trust.

Discussion
To the best of our knowledge, this research is the first study of trust and behavior toward AI-
enabled technologies across different use cases and modalities. Previous studies investigate
public opinion toward specific technologies, such as autonomous drones and vehicles, or for
some AI-applications, including autonomy in the workplace [16, 88, 89]. By contrast, our con-
joint-based study offers methodological advancements because it allows us to understand the
particular features of the AI-enabled technology that affect attitudes ranging from trust to sup-
port for its use, disconnects between the two, and variation on the basis of demographic fac-
tors. We then go further to understand the basis of public trust and support, investigating
theoretically-grounded mechanisms in a second novel study.
Our analysis shows, first, the existence of a trust paradox, wherein support for the use of
AI-enabled technologies is higher than the trust for those same technologies in certain
domains, and at certain autonomy and precision levels. The highest level of precision, with
fewer mistakes, was considerably more likely to elicit both trust and support. These findings
parallel those of public uptake of vaccines, for example, where efficacy is strongly correlated
with willingness to receive a vaccine [90].
Perceptions of controllability also play a role, in which willingness to use a particular appli-
cation increases in a mixed-initiative setting in which humans can override or interact with
the machine versus either full autonomy or full human control. Indeed, a recent report pub-
lished by the Center for Strategic and International Studies echoes this sentiment among U.S.
defense experts, stating there are “incredible new opportunities for human-machine team-
work, in which the roles of humans and machines are optimized based upon what each does
best” [91]. In light of this potential, the U.S. Air Force’s “loyal wingman” concept enables a
pilot of a manned aircraft, such as a F-22 Raptor or F-35 Lightning II, to deploy and maneuver
drones in support of mission objectives [92].
Further, the conjoint pointed to a strong preference for AI-enabled technology in the police
surveillance domain. Whereas the public was agnostic on most domains, including general
surgery, battlefield drones, social media content moderation, and autonomous vehicles, they
were substantially more supportive of AI-enabled police surveillance. This tracks with public
opinion polls showing a large plurality of adults report that they believe this technology would
be beneficial for society [93]. Populations in other countries such as Australia and the United
Kingdom have registered greater levels of skepticism, which raises the question about cross-
national variation that we suggest should be taken up by future research [94].
In terms of mechanisms, we both theorized about why individuals might trust or support
AI-enabled technologies and found that several factors play a role. These include FOMO, belief
that the benefits outweigh the risks, support for the view that the technology creates efficient
substitutes for tasks that are too dull, dirty, or dangerous for humans, and optimism about the
way that safety features are improving to reduce the potential risks imposed by emerging tech-
nologies. These attitudes are characterized by a degree of technological optimism that
improvements in innovation will provide more sustainable options over time, “an article of
faith” according to critics [95]. Further, individual attitudes are broadly consistent with an
expected utility calculation that acknowledges the risk that technology poses but expects to

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 15 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

derive some form of value from its adoption [96]. While these mediators do not constitute an
exhaustive list, they are relevant factors drawn from the literature, and future research could
investigate additional mechanisms.
Taken together, our analysis offers both theoretical and empirical insights on public atti-
tudes toward AI-enabled technologies. Beyond the trust paradox, our findings point to varia-
tion in support and trust on the basis of domain, precision, and human involvement.
Understanding the nature of public concerns and support across a range of applications and
modalities is long overdue and this research offers an initial look at how Americans consider
AI-enabled technologies. Although it is an important step in understanding public attitudes
and behaviors, as well as key factors in societal uptake of or resistance to new technologies
[97], future research should consider additional domains, such as AI in the energy sector,
manufacturing, communication [98], and politics [99] to understand additional variation
depending on the use case. Further, others could introduce the role of bias and a spectrum of
consequences to flesh out public tolerance for the range of unintended outcomes of these tech-
nologies [100]. The field of AI is rapidly evolving and research on the public uptake and resis-
tance to these technologies will have to evolve alongside those developments.

Supporting information
S1 Table. OLS table (Treatments only). Caption::Conjoint average marginal component
effects (AMCE) per attribute level. We use cars, human only autonomy, 85% precision, and
community and individual regulations as referents for domain, autonomy, precision, and reg-
ulator respectively. The dependent variable is a 5-point Likert scale.
(DOCX)
S2 Table. OLS table with controls. Caption: Conjoint average marginal component effects
(AMCE) per attribute level. We use cars, human only autonomy, 85% precision, and commu-
nity and individual regulations as referents for domain, autonomy, precision, and regulator
respectively. The dependent variable is a 5-point Likert scale. This model includes additional
levels of control variables, specifically income, education, and ethnicity.
(DOCX)
S3 Table. OLS table with interaction terms. Note: Conjoint average marginal component
(AMCE) effects per attribute level. We use cars, human only autonomy, 85% precision, and
community and individual regulations as referents for domain, autonomy, precision, and reg-
ulator respectively. The dependent variable is a 5-point Likert scale. Here we interact ethnicity
with domain and do not find strong heterogeneous treatment effects.
(DOCX)
S4 Table. Conjoint summary statistics.
(DOCX)
S5 Table. Mediation analysis summary statistics.
(DOCX)
S6 Table. Support for AI in different domains and for different purposes.
(DOCX)
S7 Table. Trust for AI in different domains and for different purposes.
(DOCX)
S8 Table. Understanding for AI in different domains and for different purposes.
(DOCX)

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 16 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

S1 File. Survey instrument.

(DOCX)
S2 File.
(DOCX)

Author Contributions
Conceptualization: Julie George.
Formal analysis: Paul Lushenko, Adi Rao.
Methodology: Paul Lushenko, Adi Rao.
Supervision: Sarah Kreps.
Writing – original draft: Julie George.
Writing – review & editing: Julie George.

References
1. Roose K. We need to talk about how good A.I. is getting [Internet]. The New York Times. 2022. Avail-
able from: https://www.nytimes.com/2022/08/24/technology/ai-technology-progress.html.
2. Horowitz MC, Kahn L. What influences attitudes about Artificial Intelligence Adoption: Evidence from
U.S. local officials. PLOS ONE. 2021Oct20; 16(10). https://doi.org/10.1371/journal.pone.0257732
PMID: 34669734
3. Morgan GM. Risk Analysis and Management. Scientific American. 1993Jul; 269(1):32–41. Available
from: https://doi.org/10.1038/scientificamerican0793-32 PMID: 8337596
4. Siegrist M, Gutscher H, Earle TC. Perception of risk: The Influence of General Trust, and general confi-
dence. Journal of Risk Research. 2006Aug15; 8(2):145–56.
5. Cobb MD, Macoubrie J. Public perceptions about nanotechnology: Risks, benefits and Trust. Journal
of Nanoparticle Research. 2004Aug; 6(4):395–405.
6. Siegrist M, Keller C, Kastenholz H, Frey S, Wiek A. Laypeople’s and experts’ perception of nanotech-
nology hazards. Risk Analysis. 2007Mar13; 27(1):59–69. https://doi.org/10.1111/j.1539-6924.2006.
00859.x PMID: 17362400
7. Frewer LJ, Scholderer J, Bredahl L. Communicating about the risks and benefits of genetically modi-
fied foods: The mediating role of Trust. Risk Analysis. 2003Dec1; 23(6):1117–33. https://doi.org/10.
1111/j.0272-4332.2003.00385.x PMID: 14641888
8. Gefen D, Karahanna E, Straub DW. Trust and tam in online shopping: An integrated model. MIS Quar-
terly. 2003Mar; 27(1):51–90.
9. Zhang B. Public opinion toward Artificial Intelligence. OSF Preprints. 2021.
10. Xiao Q, Liu H, Feldman MW. How does trust affect acceptance of a nuclear power plant (NPP): A sur-
vey among people living with Qinshan NPP in China. PLOS ONE. 2017Nov27; 12(11). https://doi.org/
10.1371/journal.pone.0187941 PMID: 29176852
11. Choi JK, Ji YG. Investigating the importance of trust on adopting an autonomous vehicle. International
Journal of Human-Computer Interaction. 2015Jul9; 31(10):692–702.
12. Luo X, Li H, Zhang J, Shim JP. Examining multi-dimensional trust and multi-faceted risk in initial
acceptance of emerging technologies: An empirical study of mobile banking services. Decision Sup-
port Systems. 2010May; 49(2):222–34.
13. Social Media Fact sheet [Internet]. Pew Research Center; 2022. Available from: https://www.
pewresearch.org/internet/fact-sheet/social-media/.
14. Gottfried J, Liedke J. Partisan divides in Media Trust Widen, driven by a decline among Republicans
[Internet]. Pew Research Center. 2021. Available from: https://www.pewresearch.org/fact-tank/2021/
08/30/partisan-divides-in-media-trust-widen-driven-by-a-decline-among-republicans/.
15. Munyaka I, Hargittai E, Redmiles E. Misinformation paradox: Older Adults are Cynical about News
Media, but Engage with It Anyway. Journal of Online Trust and Safety. 2022Sep20; 1(4).
16. Zhang Z, Jiménez FR, Cicala JE. Fear of missing out scale: A self-concept perspective. Psychology &
Marketing. 2020Sep13; 37(11):1619–34.

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 17 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

17. Paul TV, Larson DW, Wohlforth WC. Status in World Politics. New York: Cambridge University
Press; 2014.
18. Mercer J. Reputation and international politics. Ithaca: Cornell University Press; 2010.
19. Heider F. The psychology of interpersonal relations. Mansfield Center, CT: Martino Publishing; 2015.
20. Barth S, de Jong MDT. The privacy paradox–investigating discrepancies between expressed privacy
concerns and actual online behavior–A systematic literature review. Telematics and Informatics.
2017; 34(7):1038–58.
21. Pelau C, Dabija D-C, Ene I. What makes an AI device human-like? The role of interaction quality,
empathy and perceived psychological anthropomorphic characteristics in the acceptance of artificial
intelligence in the service industry. Computers in Human Behavior. 2021Sep; 122.
22. Acemoglu D, Restrepo P. Robots and Jobs: Evidence from US Labor Markets. Journal of Political
Economy. 2020Jun; 128(6):2188–244.
23. Brynjolfsson E, McAfee A. The Second Machine Age: Work, progress, and prosperity in a time of Bril-
liant Technologies [Internet]. New York City, NY: W. W. Norton & Company; 2018. Available from:
https://edisciplinas.usp.br/pluginfile.php/622156/mod_resource/content/1/Erik-Brynjolfsson-Andrew-
McAfee-Jeff-Cummings-The-Second-Machine-Age.pdf
24. DeCanio SJ. Robots and humans–complements or substitutes? Journal of Macroeconomics.
2016Sep; 49:280–91.
25. Thelen K. Varieties of capitalism: Trajectories of liberalization and the new politics of Social Solidarity.
Annual Review of Political Science. 2012Jun; 15(1):137–59.
26. Mantoux P. The Industrial Revolution in the eighteenth century: An outline of the beginnings of the
modern factory system in England. Routledge; 2015.
27. Olmstead AL, Rhode PW. Reshaping the landscape: The impact and diffusion of the tractor in Ameri-
can agriculture, 1910–1960. The Journal of Economic History. 2001; 61(3):663–98.
28. Acemoglu D, Restrepo P. Automation and new tasks: How technology displaces and Reinstates
Labor. Journal of Economic Perspectives [Internet]. 2019; 33(2):3–30. Available from: https://pubs.
aeaweb.org/doi/pdfplus/10.1257/jep.33.2.3
29. James HS Jr. The trust paradox: A survey of economic inquiries into the nature of trust and trustworthi-
ness. Journal of Economic Behavior & Organization. 2002Mar; 47(3):291–307.
30. Rainie L, Anderson J. Theme 3: Trust will not grow, but technology usage will continue to rise as a
’new normal’ sets in [Internet]. Pew Research Center; 2017. Available from: https://www.
pewresearch.org/internet/2017/08/10/theme-3-trust-will-not-grow-but-technology-usage-will-
continue-to-rise-as-a-new-normal-sets-in/.
31. Adadi A, Berrada M. Peeking inside the black-box: A survey on explainable artificial intelligence (XAI).
IEEE Access. 2018; 6:52138–60.
32. Musk E. Transparency is the key to trust [Internet]. Twitter. 2022 [cited 2022Dec20]. Available from:
https://twitter.com/elonmusk/status/1598858533608431617.
33. Putnam RD, Leonardi R, Nanetti RY. Making Democracy Work: Civic Traditions in Modern Italy. USA:
Princeton University Press; 1994.
34. Musk E. Exactly. The obvious reality, as long-time users know, is that Twitter has failed in trust &
Safety for a very long time and has interfered in elections. twitter 2.0 will be far more effective, trans-
parent and even-handed. [Internet]. Twitter. 2022 [cited 2022Dec20]. Available from: https://twitter.
com/elonmusk/status/1598004480066621441.
35. Loyola-Gonzalez O. Black-box vs. white-box: Understanding their advantages and weaknesses from
a practical point of view. IEEE Access. 2019; 7:154096–113.
36. Coppock A, McClellan OA. Validating the demographic, political, psychological, and experimental
results obtained from a new source of online survey respondents. Research & Politics. 2019; 6(1).
37. Orme BK. Getting started with Conjoint Analysis: Strategies for Product Design and Pricing Research.
Manhattan Beach, CA, USA: Research Publishers LLC; 2020.
38. Horiuchi Y, Markovich Z, Yamamoto T. Does conjoint analysis mitigate social desirability bias? Politi-
cal Analysis. 2021Sep15; 30(4):535–49.
39. Hainmueller J, Hopkins DJ, Yamamoto T. Causal inference in conjoint analysis: Understanding multi-
dimensional choices via stated preference experiments. Political Analysis. 2014; 22(1):1–30.
40. Hainmueller J, Hangartner D, Yamamoto T. Validating vignette and conjoint survey experiments
against real-world behavior. Proceedings of the National Academy of Sciences. 2015; 112(8):2395–
400. https://doi.org/10.1073/pnas.1416587112 PMID: 25646415
41. Dafoe A, Zhang B, Caughey D. Information equivalence in survey experiments. Political Analysis.
2018; 26(4):399–416.

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 18 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

42. Brutger R, Kertzer JD, Renshon J, Tingley D, Weiss CM. Abstraction and detail in experimental
design. American Journal of Political Science. 2022.
43. Suong CH, Desposato S, Gartzke E. Thinking generically and specifically in international relations sur-
vey experiments. Research & Politics. 2023Apr17; 10(2).
44. Zhang D, Mishra S, Brynjolfsson E, Etchemendy J, Ganguli D, Grosz B, et al. The AI Index 2021
Annual Report. Stanford, CA; 2021.
45. Gathering Strength, Gathering Storms: The One Hundred Year Study on Artificial Intelligence (AI100)
2021 Study Panel Report [Internet]. One Hundred Year Study on Artificial Intelligence (AI100). 2021
[cited 2022Dec20]. Available from: https://ai100.stanford.edu/gathering-strength-gathering-storms-
one-hundred-year-study-artificial-intelligence-ai100-2021-study.
46. Nam C, Walker P, Li H, Lewis M, Sycara K. Models of trust in human control of swarms with varied lev-
els of autonomy. IEEE Transactions on Human-Machine Systems. 2019; 50(3):194–204.
47. Zellers R, Holtzman A, Rashkin H, Bisk Y, Farhadi A, Roesner F, et al. Defending against neural fake
news. Neural Information Processing Systems. 2019; 812:9054–65.
48. Sigala A, Langhals B. Applications of Unmanned Aerial Systems (UAS): A Delphi study projecting
future UAS missions and relevant challenges. Drones. 2020Mar10; 4(1):8.
49. Kahn J. Can A.I. help facebook cure its disinformation problem? [Internet]. Fortune; 2021. Available
from: https://fortune.com/2021/04/06/facebook-disinformation-ai-fake-news-us-capitol-attack-social-
media-hate-speech-big-tech-solutions/.
50. Marie-Sainte SL, Saba T, Alsaleh D, Alamir Alotaibi MB. An improved strategy for predicting diagnosis,
survivability, and recurrence of breast cancer. Journal of Computational and Theoretical Nanoscience.
2019; 16(9):3705–11.
51. How the Feds coordinate with Facebook on censorship [Internet]. The Wall Street Journal. Dow Jones
& Company; 2022. Available from: https://www.wsj.com/articles/how-the-feds-coordinate-with-
facebook-twitter-white-house-social-media-emails-covid-instagram-11662761613?mod=hp_opin_
pos_1.
52. Bergstrom K, Poor N. Signaling the intent to change online communities: A case from a reddit gaming
community. Social Media + Society. 2022; 8(2):1–10.
53. Bansak K, Hainmueller J, Hopkins DJ, Yamamoto T. The number of choice tasks and survey satisfi-
cing in conjoint experiments. Political Analysis. 2018; 26(1):112–9.
54. Leeper TJ, Hobolt SB, Tilley J. Measuring subgroup preferences in conjoint experiments. Political
Analysis. 2020; 28(2):207–21.
55. Raman S, Kriner D, Ziebarth N, Simon K, Kreps S. Covid-19 booster uptake among US adults:
Assessing the impact of vaccine attributes, incentives, and context in a choice-based experiment.
Social Science & Medicine [Internet]. 2022Oct; 310. Available from: https://www.sciencedirect.com/
science/article/pii/S0277953622005834?via%3Dihub
56. Kaplan RM, Milstein A. Influence of a COVID-19 vaccine’s effectiveness and safety profile on vaccina-
tion acceptance. Proceedings of the National Academy of Sciences [Internet]. 2021; 118(10). Avail-
able from: https://doi.org/10.1073/pnas.2021726118 PMID: 33619178
57. Motta M. Can a COVID-19 vaccine live up to Americans’ expectations? A conjoint analysis of how vac-
cine characteristics influence vaccination intentions. Social Science & Medicine [Internet]. 2021; 272.
Available from: https://www.sciencedirect.com/science/article/pii/S0277953620308613?via%3Dihub
58. Castelo N, Ward AF. Conservatism predicts aversion to consequential artificial intelligence. PLOS
ONE [Internet]. 2021Dec20; 16(12). Available from: https://doi.org/10.1371/journal.pone.0261467
PMID: 34928989
59. Horowitz MC, Kahn L, Macdonald J, Schneider J. Covid-19 and public support for Autonomous Tech-
nologies—did the pandemic catalyze a world of robots? PLOS ONE [Internet]. 2022Sep28; 17(9).
Available from: https://doi.org/10.1371/journal.pone.0273941 PMID: 36170283
60. Korteling JE, van de Boer-Visschedijk GC, Blankendaal RA, Boonekamp RC, Eikelboom AR. Human-
versus Artificial Intelligence. Frontiers in Artificial Intelligence. 2021Mar25; 4. https://doi.org/10.3389/
frai.2021.622364 PMID: 33981990
61. Horowitz MC, Lin-Greenberg E. Algorithms and influence artificial intelligence and crisis decision-mak-
ing. International Studies Quarterly. 2022Oct; 66(4).
62. Imai K, Keele L, Tingley D, Yamamoto T. Unpacking the black box of causality: Learning about causal
mechanisms from experimental and observational studies. American Political Science Review. 2011;
105(4):765–89.
63. Miles C. H. On the Causal Interpretation of Randomized Interventional Indirect Effects. ArXivLabs.
Forthcoming 2022.

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 19 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

64. Chaudoin S, Gaines BJ, Livny A. Survey design, order effects, and causal mediation analysis. The
Journal of Politics. 2021Oct; 83(4):1851–6.
65. Miles CH. On the Causal Interpretation of Randomized Interventional Indirect Effects. Working Paper.
2022;
66. Simonsohn U. [103] Mediation Analysis is Counterintuitively Invalid [Internet]. Data Colada: Thinking
about evidence, and vice versa. 2022. Available from: http://datacolada.org/103
67. Imai K, Keele L, Tingley D. A general approach to causal mediation analysis. Psychological Methods.
2010Dec15; 15(4):309–34. https://doi.org/10.1037/a0020761 PMID: 20954780
68. Lin-Greenberg E. Wargame of Drones: Remotely Piloted Aircraft and Crisis Escalation. Journal of
Conflict Resolution. 2022Jun6; 66(10):1737–65.
69. Glazier RA, Boydstun AE, Feezell JT. Self-coding: A method to assess semantic validity and bias
when coding open-ended responses. Research & Politics. 2021Jul27; 8(3).
70. Christensen DM, Manley J, Resendez J. Medical algorithms are failing communities of color [Internet].
Health Affairs Forefront. 2021. Available from: https://www.healthaffairs.org/do/10.1377/forefront.
20210903.976632/full/.
71. Osoba O, Welser W. An intelligence in our image: The risks of bias and errors in artificial intelligence.
Santa Monica, CA: RAND Corporation; 2017.
72. Bruyn AD, Viswanathan V, Beh YS, Brock JK-U, Wangenheim Fvon. Artificial Intelligence and Market-
ing: Pitfalls and opportunities. Journal of Interactive Marketing. 2020Aug; 51:91–105.
73. Zhang B. Public opinion lessons for AI regulation [Internet]. Brookings. 2019. Available from: https://
www.brookings.edu/research/public-opinion-lessons-for-ai-regulation.
74. Vogels EA. Millennials stand out for their technology use, but older generations also Embrace Digital
Life [Internet]. Pew Research Center; 2019. Available from: https://www.pewresearch.org/fact-tank/
2019/09/09/us-generations-technology-use/.
75. Gelles-Watnick R. U.S. women more concerned than men about some AI developments, especially
driverless cars [Internet]. Pew Research Center. 2022 [cited 2022Dec20]. Available from: https://
www.pewresearch.org/fact-tank/2022/08/03/u-s-women-more-concerned-than-men-about-some-ai-
developments-especially-driverless-cars/.
76. Castro D. Many women aren’t sold on AI. that’s a problem. [Internet]. Center for Data Innovation.
2019 [cited 2022Dec20]. Available from: https://datainnovation.org/2019/03/many-women-arent-sold-
on-ai-thats-a-problem/
77. Bleicher A. Demystifying the Black Box That Is AI [Internet]. Scientific American. 2017. Available
from: https://www.scientificamerican.com/article/demystifying-the-black-box-that-is-ai/
78. Kreps SE, Wallace GPR. International law, military effectiveness, and public support for drone strikes.
Journal of Peace Research. 2016Sep25; 53(6):830–44.
79. Macdonald J, Schneider J. Battlefield responses to new technologies: Views from the ground on
Unmanned Aircraft. Security Studies [Internet]. 2019Feb7; 28(2):216–49. Available from: https://doi.
org/10.1080/09636412.2019.1551565
80. Strawser BJ, editor. Killing by remote control: The ethics of an unmanned military. Oxford University
Press; 2013.
81. Smith A, Anderson M. 3. Americans’ attitudes toward driverless vehicles [Internet]. Pew Research
Center: Internet, Science & Tech. 2017 [cited 2022Dec20]. Available from: https://www.pewresearch.
org/internet/2017/10/04/americans-attitudes-toward-driverless-vehicles/
82. Bartlett JS. How Much Automation Does Your Car Really Have? [Internet]. Consumer Reports. 2021
[cited 2022Dec20]. Available from: https://www.consumerreports.org/automotive-technology/how-
much-automation-does-your-car-really-have-level-2-a3543419955/.
83. Gorwa R, Binns R, Katzenbach C. Algorithmic content moderation: Technical and political challenges
in the automation of platform governance. Big Data & Society. 2020Feb28; 7(1).
84. Watts T. “I Need Your Clothes, Your Boots and Your Terminator Tropes”: A Primer on the Terminator
Franchise [Internet]. AutoNorms. 2021 [cited 2022Dec20]. Available from: https://www.autonorms.eu/
i-need-your-clothes-your-boots-and-your-terminator-tropes-the-popular-depiction-of-killer-robots-in-
american-culture/.
85. Baron RM, Kenny DA. The moderator–mediator variable distinction in social psychological research:
Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology.
1986; 51(6):1173–82. https://doi.org/10.1037//0022-3514.51.6.1173 PMID: 3806354
86. Tomz MR, Weeks JLP. Human Rights and Public Support for War. Journal of Politics. 2020Jan; 82
(1):182–94.

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 20 / 21

PLOS ONE Exploring the artificial intelligence “Trust paradox”: Evidence from a survey experiment in the United States

87. Fisk K, Merolla JL, Ramos JM. Emotions, terrorist threat, and drones: Anger drives support for drone
strikes. Journal of Conflict Resolution. 2018May3; 63(4):976–1000.
88. Horowitz MC. Public opinion and the politics of the killer robots debate. Research and Politics. 2016; 3
(1):1–8.
89. Othman K. Public acceptance and perception of Autonomous Vehicles: A comprehensive review. AI
and Ethics. 2021; 1(3):355–87. https://doi.org/10.1007/s43681-021-00041-8 PMID: 34790943
90. Kreps S, Prasad S, Brownstein JS, Hswen Y, Garibaldi BT, Zhang B, et al. Factors associated with us
adults’ likelihood of accepting COVID-19 vaccination. JAMA Network Open. 2020; 3(10):1–13. https://
doi.org/10.1001/jamanetworkopen.2020.25594 PMID: 33079199
91. Mulchandani N, Shanahan J. Software-defined warfare: Architecting the DOD’s transition to the Digital
age [Internet]. Center for Strategic and International Studies. 2022. Available from: https://www.csis.
org/analysis/software-defined-warfare-architecting-dods-transition-digital-age.
92. Fish T. Uncrewed ambitions of the Loyal Wingman [Internet]. Airforce Technology. 2022 [cited
2022Dec20]. Available from: https://www.airforce-technology.com/features/uncrewed-ambitions-of-
the-loyal-wingman/.
93. Rainie L, Funk C, Anderson M, Tyson A. 4. Americans cautious about the deployment of driverless
cars [Internet]. Pew Research Center; 2022. Available from: https://www.pewresearch.org/internet/
2022/03/17/americans-cautious-about-the-deployment-of-driverless-cars/.
94. Ritchie KL, Cartledge C, Growns B, Yan A, Wang Y, Guo K, et al. Public attitudes towards the use of
automatic facial recognition technology in criminal justice systems around the world. PLOS ONE.
2021; 16(10). https://doi.org/10.1371/journal.pone.0258241 PMID: 34644306
95. Basiago AD. The limits of technological optimism. The Environmentalist. 1994; 14(1):17–22.
96. Sparks P, Shepherd R, Frewer LJ. Assessing and structuring attitudes toward the use of gene technol-
ogy in food production: The role of perceived ethical obligation. Basic and Applied Social Psychology.
1995; 16(3):267–85.
97. Juma C. Innovation and its enemies: Why people resist new technologies. New York: Oxford Univer-
sity Press; 2019.
98. Jakesch M, French M, Ma X, Hancock JT, Naaman M. AI-Mediated Communication:: How the Percep-
tion that Profile Text was Written by AI Affects Trustworthiness. Proceedings of the 2019 CHI Confer-
ence on Human Factors in Computing Systems. 2019May:1–13.
99. Kreps SE. Social Media and International Relations. Cambridge, United Kingdom: Cambridge Univer-
sity Press; 2020.
100. Challen R, Denny J, Pitt M, Gompels L, Edwards T, Tsaneva-Atanasova K. Artificial Intelligence, Bias
and clinical safety. BMJ Quality & Safety. 2019; 28(3):231–7. https://doi.org/10.1136/bmjqs-2018-
008370 PMID: 30636200

PLOS ONE | https://doi.org/10.1371/journal.pone.0288109 July 18, 2023 21 / 21