Related Literature Group 4, design: Difference between revisions

From Control Systems Technology Group
Jump to navigation Jump to search
No edit summary
No edit summary
Line 15: Line 15:
* We found that the designed characteristics of VAs affects some aspects of user impressions (i.e. personality and trustworthiness) of the VA, while other impressions are not affected (i.e. social ability). We also found that gender matching between the agent and the user affect user impressions.  // Gender similarity --> more trustworthy <ref name="akbar2018">Akbar, F., Grover, T., Mark, G., & Zhou, M. X. (2018, March 5). The effects of virtual agents’ characteristics on user impressions and language use. International Conference on Intelligent User Interfaces, Proceedings IUI. https://doi.org/10.1145/3180308.3180365</ref>
* We found that the designed characteristics of VAs affects some aspects of user impressions (i.e. personality and trustworthiness) of the VA, while other impressions are not affected (i.e. social ability). We also found that gender matching between the agent and the user affect user impressions.  // Gender similarity --> more trustworthy <ref name="akbar2018">Akbar, F., Grover, T., Mark, G., & Zhou, M. X. (2018, March 5). The effects of virtual agents’ characteristics on user impressions and language use. International Conference on Intelligent User Interfaces, Proceedings IUI. https://doi.org/10.1145/3180308.3180365</ref>


====Human-likeness====
* In addition, the similarity of an agent to the learner positively influences the learner’s motivation (Bailenson, Blascovich, & Guadagno, 2008 in a study with undergraduates). For example, computer-based female agents yielded better motivational outcomes for undergraduate women if they matched the students with respect to race and gender (Rosenberg-Kima, Plant, Doerr, & Baylor, 2010). Another study, conducted with undergraduates by Rosenberg-Kima, Baylor, Plant, and Doerr (2008) revealed that a female agent rated as young, attractive and “cool” succeeded in enhancing young female students’ self-efficacy, which is believed to be a driving force behind motivation (Bandura, 1997). All these findings are theoretically supported by Bandura’s social cognitive learning theory which states that people often learn behavior and norms by imitating people whom they perceive as similar (or superior: higher in rank or status) to them and who are therefore rather accepted as social role models (Bandura, 1986). This finding is supported by another study of Gulz, Haake, and Tärning (2007) which demonstrated that participants prefer same-gender agents when they are asked to choose their preferred agent as presenter for a multimedia slideshow. <ref name="shiban2015"/>
 
====Appearance and human-likeness====


* In this study, it was noted that PIAs’ human-like features could influence the manner in which participants (i.e. males or females) related to a particular PIA and participants’ preference level of a particular PIA. The result of the Fischer’s exact test conducted in this study (Table V) found no statistically significant effect between participants’ gender (male or female) and their preferred gender of PIAs. It is concluded that the gender of participants (i.e. male or female) has no role (impact) on their preferences regarding the type, features, or gender of PIA, neither on their preference level of PIA. <ref name="mabanza2019">Mabanza, N. (2019). Gender influences on preference of pedagogical interface agents. 2018 International Conference on Intelligent and Innovative Computing Applications, ICONIC 2018. https://doi.org/10.1109/ICONIC.2018.8601292</ref>
* In this study, it was noted that PIAs’ human-like features could influence the manner in which participants (i.e. males or females) related to a particular PIA and participants’ preference level of a particular PIA. The result of the Fischer’s exact test conducted in this study (Table V) found no statistically significant effect between participants’ gender (male or female) and their preferred gender of PIAs. It is concluded that the gender of participants (i.e. male or female) has no role (impact) on their preferences regarding the type, features, or gender of PIA, neither on their preference level of PIA. <ref name="mabanza2019">Mabanza, N. (2019). Gender influences on preference of pedagogical interface agents. 2018 International Conference on Intelligent and Innovative Computing Applications, ICONIC 2018. https://doi.org/10.1109/ICONIC.2018.8601292</ref>
Line 33: Line 35:


* In addition to understanding human social behavior around computers, another extensive line of work examines humans’ responses to computers with more expressive, human-like qualities, such as faces and facial expressions. In general, these studies have found that anthropomorphic properties of computers influence users’ perceptions (e.g., [13, 70, 86]), attitude (e.g., [13]), and behavior around such systems (e.g., [48, 86, 105]). For example, Sproull et al. [86] report that people respond to a text-based interface differently than to a talking face. On the one hand, users are aroused more and present themselves more positively when interacting with a talking face. On the other hand, users are less relaxed or assured when interacting with a talking face. Another study shows that users find the interface with the anthropomorphic qualities—faces and facial expressions—more likeable and engaging, although such an interface takes the users’ effort to interpret the meaning of the human-like expressions and may even be a distraction [48, 86]. More recent studies show that human-like features with higher realism elicit more positive social interactions while having no significant impact on user task performance [105]. Furthermore, a study reveals that anthropomorphism may even elicit user objections due to users’ own biases (e.g., sexism) [70]. <ref name="zhou2019">Zhou, M. X., Yang, H., Mark, G., & Li, J. (2019). Trusting Virtual Agents: The Effect of Personality. ACM Trans. Interact. Intell. Syst, 9(3). https://doi.org/10.1145/3232077</ref>
* In addition to understanding human social behavior around computers, another extensive line of work examines humans’ responses to computers with more expressive, human-like qualities, such as faces and facial expressions. In general, these studies have found that anthropomorphic properties of computers influence users’ perceptions (e.g., [13, 70, 86]), attitude (e.g., [13]), and behavior around such systems (e.g., [48, 86, 105]). For example, Sproull et al. [86] report that people respond to a text-based interface differently than to a talking face. On the one hand, users are aroused more and present themselves more positively when interacting with a talking face. On the other hand, users are less relaxed or assured when interacting with a talking face. Another study shows that users find the interface with the anthropomorphic qualities—faces and facial expressions—more likeable and engaging, although such an interface takes the users’ effort to interpret the meaning of the human-like expressions and may even be a distraction [48, 86]. More recent studies show that human-like features with higher realism elicit more positive social interactions while having no significant impact on user task performance [105]. Furthermore, a study reveals that anthropomorphism may even elicit user objections due to users’ own biases (e.g., sexism) [70]. <ref name="zhou2019">Zhou, M. X., Yang, H., Mark, G., & Li, J. (2019). Trusting Virtual Agents: The Effect of Personality. ACM Trans. Interact. Intell. Syst, 9(3). https://doi.org/10.1145/3232077</ref>
* Physically present agent leads to better motivational outcomes than a voice or text-box. It is important to design agents realistically, because e.g. cartoon figures have been shown to diminish the positive motivating effects in comparison to realistic figures. <ref name="shiban2015"/>
* It was found also that attractive VT did not improve perceived value of exercise and perceived risks of health. This result is different from some prior studies (e.g. Shiban et al., 2015) which find that attractive virtual pedagogical agents are effective for engaging students in learning. There are two possible explanations for this observed difference. First, attractive virtual agents may catch the attention of users and make users more confident in using a VTS. However, they may also distract users from core learning materials (Moreno and Flowerday, 2006). Second, attractive agents are closer to friends rather than experts. The incongruence of expert perception may lead to contextual irrelevance which makes users take health-related information from virtual trainers lightly (Veletsianos, 2010). <ref name="kwok2021">Kwok, R. C. W., Leung, A. C. M., Hui, S. S. chuen, & Wong, C. C. K. (2021). Virtual trainer system: a tool to increase exercise participation and work productivity. Internet Research. https://doi.org/10.1108/INTR-04-2020-0236</ref>
* If McCloud’s (1993) framework is applied, a teacher character, representing the other to a higher extent than a learning companion, might benefit from more realism in the representation. A learning companion character, being to a higher extent conceived of as an extension of oneself, may, on the other hand, benefit from a more iconic representation. <ref name="gulz2006">Gulz, A., & Haake, M. (2006). Design of animated pedagogical agents - A look at their look. International Journal of Human Computer Studies, 64(4), 322–339. https://doi.org/10.1016/j.ijhcs.2005.08.006</ref>
* The findings of this experiment suggest a new design method for human-agent collaboration work. If we use an agent having a non-human-like appearance (for example, the robot-like agent), the user seemed to attribute much responsibility to the agent and not to trust the agent. We suggest using the agent having human-like appearance in human-agent collaboration work from this experiment. A question is raised about the commonly accepted beliefs about agent design for social tasks. We should consider the attribution of responsibility to construct trustworthy agents. <ref name="matsui2021">Matsui, T., & Koike, A. (2021). Who is to blame? The appearance of virtual agents and the attribution of perceived responsibility. Sensors, 21(8), 1–13. https://doi.org/10.3390/s21082646</ref>
====Behaviour and personality====


===References===
===References===


<references/>
<references/>

Revision as of 19:21, 9 May 2021

Excerpts & citations

Gender

  • Females chose to gender-match and to interact with a more realistic VA. Males exhibited little preference for either gender, and a greater preference than females for realistic VAs. Thus, where it is not feasible to gender-match in SSCO, the recommendation is to implement a realistic female VA. [1]
  • Young and attractive female agent positively impacts interest in learning, an older an unattractive male agent does not impact motivation. [2]
  • As research suggests that the combination of an agent’s gender and personality can play an important role in user perceptions and expectations [1, 37], employing a male agent instead may result in some significant differences in user perceptions or ratings of the agent. We encourage future work to investigate how gender and personality of a workplace productivity agent might influence user experience. [3]
  • participants prefer same-gender agents when they are asked to choose their preferred agent as presenter for a multimedia slideshow. [2]
  • We found that the designed characteristics of VAs affects some aspects of user impressions (i.e. personality and trustworthiness) of the VA, while other impressions are not affected (i.e. social ability). We also found that gender matching between the agent and the user affect user impressions. // Gender similarity --> more trustworthy [4]
  • In addition, the similarity of an agent to the learner positively influences the learner’s motivation (Bailenson, Blascovich, & Guadagno, 2008 in a study with undergraduates). For example, computer-based female agents yielded better motivational outcomes for undergraduate women if they matched the students with respect to race and gender (Rosenberg-Kima, Plant, Doerr, & Baylor, 2010). Another study, conducted with undergraduates by Rosenberg-Kima, Baylor, Plant, and Doerr (2008) revealed that a female agent rated as young, attractive and “cool” succeeded in enhancing young female students’ self-efficacy, which is believed to be a driving force behind motivation (Bandura, 1997). All these findings are theoretically supported by Bandura’s social cognitive learning theory which states that people often learn behavior and norms by imitating people whom they perceive as similar (or superior: higher in rank or status) to them and who are therefore rather accepted as social role models (Bandura, 1986). This finding is supported by another study of Gulz, Haake, and Tärning (2007) which demonstrated that participants prefer same-gender agents when they are asked to choose their preferred agent as presenter for a multimedia slideshow. [2]

Appearance and human-likeness

  • In this study, it was noted that PIAs’ human-like features could influence the manner in which participants (i.e. males or females) related to a particular PIA and participants’ preference level of a particular PIA. The result of the Fischer’s exact test conducted in this study (Table V) found no statistically significant effect between participants’ gender (male or female) and their preferred gender of PIAs. It is concluded that the gender of participants (i.e. male or female) has no role (impact) on their preferences regarding the type, features, or gender of PIA, neither on their preference level of PIA. [5]
  • On the other hand, previous studies have also shown that developing overly humanized agents results in high expectations and uncanny feelings. [6]
  • Many participants saw the human-like appearance of the VA prototype as setting the wrong expectations in terms of its capabilities, and they were disappointed when the agent’s intelligence only extended towards responding to the their emotion and prompting more self-reflection. (…) We believe incorporating human-like qualities and emotional intelligence into future agents to be worthwhile; however, intelligence should also extend into other aspects of the agent’s capabilities in order to better help users be as efficient as possible in achieving their work goals. [3]
  • A study by Go and Sundar [7] states that revealing the identity of a chatbot as a non-human can have a positive effect: user will have less high expectations about the conversation, and will be impressed when an agent shows human-like behaviour. Furthermore, they emphasize the importance of the conversational style between a human an a computer. When the dialogue resembles that of an actual human, perceived feelings of social presence and homophily will increase, leading to more positive attitudes towards the agent (and in turn potential desired behaviour consequences).
  • the presence of a representation produced more positive social interactions than not having a representation [8]
    • human-like representations with higher realism produced more positive social interactions than representations with lower realism; however, this effect was only found when subjective measures were used. Behavioral measures did not reveal a significant difference between representations of low and high realism.
      • the difference we found may also be driven by demand characteristics. Participants interacting with an animated character (as opposed to a photograph) may suppose that the researcher is expecting a high appraisal.
    • while the presence of a face is better than no face at all, the quality of the face matters much less.
      • it is quite possible that animating highly realistic faces inherently allows for residual attributes of the faces that are negative—for example making 3D human faces may produce gestures and animations that appear unnatural or disturbing
    • while most studies have found that interface agents have positive effects on task performance, these effects are overall actually quite small.
  • In addition to understanding human social behavior around computers, another extensive line of work examines humans’ responses to computers with more expressive, human-like qualities, such as faces and facial expressions. In general, these studies have found that anthropomorphic properties of computers influence users’ perceptions (e.g., [13, 70, 86]), attitude (e.g., [13]), and behavior around such systems (e.g., [48, 86, 105]). For example, Sproull et al. [86] report that people respond to a text-based interface differently than to a talking face. On the one hand, users are aroused more and present themselves more positively when interacting with a talking face. On the other hand, users are less relaxed or assured when interacting with a talking face. Another study shows that users find the interface with the anthropomorphic qualities—faces and facial expressions—more likeable and engaging, although such an interface takes the users’ effort to interpret the meaning of the human-like expressions and may even be a distraction [48, 86]. More recent studies show that human-like features with higher realism elicit more positive social interactions while having no significant impact on user task performance [105]. Furthermore, a study reveals that anthropomorphism may even elicit user objections due to users’ own biases (e.g., sexism) [70]. [9]
  • Physically present agent leads to better motivational outcomes than a voice or text-box. It is important to design agents realistically, because e.g. cartoon figures have been shown to diminish the positive motivating effects in comparison to realistic figures. [2]
  • It was found also that attractive VT did not improve perceived value of exercise and perceived risks of health. This result is different from some prior studies (e.g. Shiban et al., 2015) which find that attractive virtual pedagogical agents are effective for engaging students in learning. There are two possible explanations for this observed difference. First, attractive virtual agents may catch the attention of users and make users more confident in using a VTS. However, they may also distract users from core learning materials (Moreno and Flowerday, 2006). Second, attractive agents are closer to friends rather than experts. The incongruence of expert perception may lead to contextual irrelevance which makes users take health-related information from virtual trainers lightly (Veletsianos, 2010). [10]
  • If McCloud’s (1993) framework is applied, a teacher character, representing the other to a higher extent than a learning companion, might benefit from more realism in the representation. A learning companion character, being to a higher extent conceived of as an extension of oneself, may, on the other hand, benefit from a more iconic representation. [11]
  • The findings of this experiment suggest a new design method for human-agent collaboration work. If we use an agent having a non-human-like appearance (for example, the robot-like agent), the user seemed to attribute much responsibility to the agent and not to trust the agent. We suggest using the agent having human-like appearance in human-agent collaboration work from this experiment. A question is raised about the commonly accepted beliefs about agent design for social tasks. We should consider the attribution of responsibility to construct trustworthy agents. [12]

Behaviour and personality

References

  1. Payne, J., Szymkowiak, A., Robertson, P., & Johnson, G. (2013). Gendering the machine: Preferred virtual assistant gender and realism in self-service. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8108 LNAI, 106–115. https://doi.org/10.1007/978-3-642-40415-3_9
  2. 2.0 2.1 2.2 2.3 Shiban, Y., Schelhorn, I., Jobst, V., Hörnlein, A., Puppe, F., Pauli, P., & Mühlberger, A. (2015). The appearance effect: Influences of virtual agent features on performance and motivation. Computers in Human Behavior, 49, 5–11. https://doi.org/10.1016/j.chb.2015.01.077
  3. 3.0 3.1 Grover, T., Rowan, K., Suh, J., McDuff, D., & Czerwinski, M. (2020). Design and evaluation of intelligent agent prototypes for assistance with focus and productivity at work. International Conference on Intelligent User Interfaces, Proceedings IUI, 20, 390–400. https://doi.org/10.1145/3377325.3377507
  4. Akbar, F., Grover, T., Mark, G., & Zhou, M. X. (2018, March 5). The effects of virtual agents’ characteristics on user impressions and language use. International Conference on Intelligent User Interfaces, Proceedings IUI. https://doi.org/10.1145/3180308.3180365
  5. Mabanza, N. (2019). Gender influences on preference of pedagogical interface agents. 2018 International Conference on Intelligent and Innovative Computing Applications, ICONIC 2018. https://doi.org/10.1109/ICONIC.2018.8601292
  6. Chaves, A. P., & Gerosa, M. A. (2021). How Should My Chatbot Interact? A Survey on Social Characteristics in Human–Chatbot Interaction Desi | Enhanced Reader. International Journal of Human–Computer Interaction, 37(8), 729–758. https://doi.org/10.1080/10447318.2020.1841438
  7. Go, E., & Sundar, S. S. (2019). Humanizing chatbots: The effects of visual, identity and conversational cues on humanness perceptions. Computers in Human Behavior, 97, 304–316. https://doi.org/10.1016/j.chb.2019.01.020
  8. Yee, N., Bailenson, J. N., & Rickertsen, K. (2007). A Meta-Analysis of the Impact of the Inclusion and Realism of Human-Like Faces on User Experiences in Interfaces.
  9. Zhou, M. X., Yang, H., Mark, G., & Li, J. (2019). Trusting Virtual Agents: The Effect of Personality. ACM Trans. Interact. Intell. Syst, 9(3). https://doi.org/10.1145/3232077
  10. Kwok, R. C. W., Leung, A. C. M., Hui, S. S. chuen, & Wong, C. C. K. (2021). Virtual trainer system: a tool to increase exercise participation and work productivity. Internet Research. https://doi.org/10.1108/INTR-04-2020-0236
  11. Gulz, A., & Haake, M. (2006). Design of animated pedagogical agents - A look at their look. International Journal of Human Computer Studies, 64(4), 322–339. https://doi.org/10.1016/j.ijhcs.2005.08.006
  12. Matsui, T., & Koike, A. (2021). Who is to blame? The appearance of virtual agents and the attribution of perceived responsibility. Sensors, 21(8), 1–13. https://doi.org/10.3390/s21082646