Karl F. MacDorman

Revisiting the uncanny valley theory: Developing and validating an alternative to the Godspeed indices

Mori (1970) proposed a hypothetical graph describing a nonlinear relation between a characters degree of human likeness and the emotional response of the human perceiver. However, the index construction of these variables could result in their strong correlation, thus preventing rated characters from being plotted accurately. Phase 1 of this study tested the indices of the Godspeed questionnaire as measures of humanlike characters. The results indicate significant and strong correlations among the relevant indices (Bartneck, Kulic, Croft, & Zoghbi, 2009). Phase 2 of this study developed alternative indices with nonsignificant correlations (p>.05) between the proposed y-axis eeriness and x-axis perceived humanness (r=.02). The new humanness and eeriness indices facilitate plotting relations among rated characters of varying human likeness.

Does Japan really have robot mania? Comparing attitudes by implicit and explicit measures

Japan has more robots than any other country with robots contributing to many areas of society, including manufacturing, healthcare, and entertainment. However, few studies have examined Japanese attitudes toward robots, and none has used implicit measures. This study compares attitudes among the faculty of a US and a Japanese university. Although the Japanese faculty reported many more experiences with robots, implicit measures indicated both faculties had more pleasant associations with humans. In addition, although the US faculty reported people were more threatening than robots, implicit measures indicated both faculties associated weapons more strongly with robots than with humans. Despite the media’s hype about Japan’s robot ‘craze,’ response similarities suggest factors other than attitude better explain robot adoption. These include differences in history and religion, personal and human identity, economic structure, professional specialization, and government policy. Japanese robotics offers a unique reference from which other nations may learn.

Sensitivity to the proportions of faces that vary in human likeness

Despite the often quoted adage beauty is in the eye of the beholder, studies indicate people perceive certain facial and bodily proportions as attractive regardless of their culture. This preference, which is present even in infants, may be more hardwired than learned. Designers of computer games, animation, virtual reality, and robots must make choices about how to depict humanlike forms. An understanding of human perception and preferences can lead to design principles for successful interaction. This study measured human responses to varying facial proportions in people, androids, mechanical-looking robots, and two- and three-dimensional characters. Participants showed greater agreement on the best proportions of faces they considered more humanlike and more attractive and less tolerance for deviation from these proportions in more attractive faces.

Reducing consistency in human realism increases the uncanny valley effect; increasing category uncertainty does not

Human replicas may elicit unintended cold, eerie feelings in viewers, an effect known as the uncanny valley. Masahiro Mori, who proposed the effect in 1970, attributed it to inconsistencies in the replicas realism with some of its features perceived as human and others as nonhuman. This study aims to determine whether reducing realism consistency in visual features increases the uncanny valley effect. In three rounds of experiments, 548 participants categorized and rated humans, animals, and objects that varied from computer animated to real. Two sets of features were manipulated to reduce realism consistency. (For humans, the sets were eyes-eyelashes-mouth and skin-nose-eyebrows.) Reducing realism consistency caused humans and animals, but not objects, to appear eerier and colder. However, the predictions of a competing theory, proposed by Ernst Jentsch in 1906, were not supported: The most ambiguous representations-those eliciting the greatest category uncertainty-were neither the eeriest nor the coldest.

Does social desirability bias favor humans? Explicit-implicit evaluations of synthesized speech support a new HCI model of impression management

Do people treat computers as social actors? To answer this question, researchers have measured the extent to which computers elicit social responses in people, such as impression management strategies for influencing the perceptions of others. But on this question findings in the literature conflict. To make sense of these findings, the present study proposes a dual-process model of impression management in human-computer interaction. The model predicts that, although machines may elicit nonconscious impression management strategies, they do not generally elicit conscious impression management strategies. One such strategy is presenting oneself favorably to others, which can be measured as social desirability bias when comparing self-reported preferences with implicit preferences. The current study uses both a questionnaire and an implicit association test (IAT) to compare attitudes toward human and machine speech. Although past studies on social desirability bias have demonstrated peoples tendency to underreport their preference for the preferred group when comparing two human groups, the current study found that, when comparing human speech and machine-synthesized speech, participants instead overreported their preference for the preferred (human) group. This finding supports the proposed dual-process model of impression management, because participants did not consciously treat computers as social actors.

The uncanny valley does not interfere with level 1 visual perspective taking

When a computer-animated human character looks eerily realistic, viewers report a loss of empathy; they have difficulty taking the characters perspective. To explain this perspective-taking impairment, known as the uncanny valley, a novel theory is proposed: The more human or less eerie a character looks, the more it interferes with level 1 visual perspective taking when the characters perspective differs from that of the human observer (e.g., because the character competitively activates shared circuits in the observers brain). The proposed theory is evaluated in three experiments involving a dot-counting task in which participants either assumed or ignored the perspective of characters varying in their human photorealism and eeriness. Although response times and error rates were lower when the number of dots faced by the observer and character were the same (congruent condition) than when they were different (incongruent condition), no consistent pattern emerged between the human photorealism or eeriness of the characters and participants response times and error rates. Thus, the proposed theory is unsupported for level 1 visual perspective taking. As the effects of the uncanny valley on empathy have not previously been investigated systematically, these results provide evidence to eliminate one possible explanation.

An Improved Usability Measure Based on Novice and Expert Performance

The novice–expert ratio method (NEM) pinpoints user interface design problems by identifying the steps in a task that have a high ratio of novice to expert completion time. This study tested the construct validity of NEMs ratio measure against common alternatives. Data were collected from 337 participants who separately performed 10 word-completion tasks on a cellular phone interface. The logarithm, ratio, Cohens d, and Hedgess ĝ measures had similar construct validity, but Hedgess ĝ provided the most accurate measure of effect size. All these measures correlated more strongly with self-reported interface usability and interface knowledge when applied to the number of actions required to complete a task than when applied to task completion time. A weighted average of both measures had the highest correlation. The relatively high correlation between self-reported interface usability and a weighted Hedgess ĝ measure as compared to the correlations found in the literature indicates the usefulness of the weighted Hedgess ĝ measure in identifying usability problems.

THE INFLUENCE OF HOLISTIC AND ANALYTIC COGNITIVE STYLES ON ONLINE INFORMATION DESIGN: Toward a communication theory of cultural cognitive design

Although studies have linked culture to online user preferences and performance, few communication researchers have recognized the impact of culture on online information design and usability. It is important to ask if people are better able to use and prefer Web sites created by designers from their own culture. We propose that to improve computer-mediated communication, Web site design should accommodate culturally diverse user groups. First, a body of research is presented that aligns East Asian cultures with more holistic cognitive styles and Western cultures with more analytical cognitive styles. Building on this contrast, a theory of cultural cognitive design is proposed as a means of understanding how cognitive styles that develop under the influence of culture lead to different ways of designing and organizing information for the Web.

The Aesthetic Dimensions of U.S. and South Korean Responses to Web Home Pages: A Cross-Cultural Comparison

Culturally influenced preferences in website aesthetics is a topic often neglected by scholars in human–computer interaction. Kim, Lee, and Choi (2003) identified aesthetic design factors of web home pages that elicited particular responses in South Korean web users based on 13 secondary emotional dimensions. This study extends Kim et al.s work to U.S. participants, comparing the original South Korean findings with U.S. findings. Results show that U.S. participants reliably applied translations of the emotional adjectives used in the South Korean study to the home pages. However, factor analysis revealed that the aesthetic perceptions of U.S. and South Korean participants formed different aesthetic dimensions composed of different sets of emotional adjectives, suggesting that U.S. and South Korean people perceive the aesthetics of home pages differently. These results indicate that website aesthetics can vary significantly between cultures.

Familiar faces rendered strange: Why inconsistent realism drives characters into the uncanny valley.

Computer-modeled characters resembling real people sometimes elicit cold, eerie feelings. This effect, called the uncanny valley, has been attributed to uncertainty about whether the character is human or living or real. Uncertainty, however, neither explains why anthropomorphic characters lie in the uncanny valley nor their characteristic eeriness. We propose that realism inconsistency causes anthropomorphic characters to appear unfamiliar, despite their physical similarity to real people, owing to perceptual narrowing. We further propose that their unfamiliar, fake appearance elicits cold, eerie feelings, motivating threat avoidance. In our experiment, 365 participants categorized and rated objects, animals, and humans whose realism was manipulated along consistency-reduced and control transitions. These data were used to quantify a Bayesian model of categorical perception. In hypothesis testing, we found reducing realism consistency did not make objects appear less familiar, but only animals and humans, thereby eliciting cold, eerie feelings. Next, structural equation models elucidated the relation among realism inconsistency (measured objectively in a two-dimensional Morlet wavelet domain inspired by the primary visual cortex), realism, familiarity, eeriness, and warmth. The fact that reducing realism consistency only elicited cold, eerie feelings toward anthropomorphic characters, and only when it lessened familiarity, indicates the role of perceptual narrowing in the uncanny valley.