"I don’t like my avatar": Investigating Human Digital Doubles CCS: Computing methodologies Animation CCS: Human-centered computing Empirical studies in HCI

Siyi Liu, Kazi Injamamul Haque, Zerrin Yumak · 2025

Read "I don’t like my avatar": Investigating Human Digital Doubles CCS: Computing methodologies Animation CCS: Human-centered computing Empirical studies in HCI by Siyi Liu, Kazi Injamamul Haque, Zerrin Yumak (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-17748.txt.

Introduction

Siyi Liu, Kazi Injamamul Haque, and Zerrin Yumak ask people how they feel about their own digital doubles. Their 2025 arXiv paper notes that avatars of real people are now easy to make on consumer gear. The team tested how avatar style and familiarity shape the bond: realistic or cartoon, and a face that is one's own, an acquaintance's, or a stranger's.

They built both styles, realistic MetaHumans and cartoon ReadyPlayerMe figures, with facial animation from performance capture. In a controlled study, more realistic looks raised identification, perceived realism, and social presence. Familiar faces cut the other way, above all when realistic: people liked their own double least of all. Participants still knew the double was theirs. They judged their own avatar harshly yet forgave the doubles of acquaintances and strangers. The title quotes the mood: I don't like my avatar.

Catalog record

CREATORSiyi Liu, Kazi Injamamul Haque, Zerrin Yumak
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-17748

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

"I don’t like my avatar": Investigating Human Digital Doubles CCS: Computing methodologies Animation CCS: Human-centered computing Empirical studies in HCI Siyi Liu, Kazi Injamamul Haque, Zerrin Yumak Abstract . Creating human digital doubles is becoming easier and much more accessible to everyone using consumer grade devices. In this work, we investigate how avatar style (realistic vs cartoon) and avatar familiarity (self, acquaintance, unknown person) affect self/other-identification, perceived realism, affinity and social presence with a controlled offline experiment. We created two styles of avatars (realistic-looking MetaHumans and cartoon-looking ReadyPlayerMe avatars) and facial animations stimuli for them using performance capture. Questionnaire responses demonstrate that higher appearance realism leads to a higher level of identification, perceived realism and social presence. However, avatars with familiar faces, especially those with high appearance realism, lead to a lower level of identification, perceived realism, and affinity. Although participants identified their digital doubles as their own, they consistently did not like their avatars, especially of realistic appearance. But they were less critical and more forgiving about their acquaintance’s or an unknown person’s digital double. Figure 1 . We investigate the perception of human digital doubles with realistic (top row) and stylized (bottom row) appearances and three levels of familiarity - self, acquaintance and unknown person. Teaser figure containing examples of human digital doubles. 1. Introduction We have seen game creators avidly using realistic digital doubles of actors/actresses in video games in recent years. Keanu Reeves’ digital double Johnny Silverhand appears in Cyberpunk 2077 while Gina Torres appears as Kirkan in Immortals of the Aveum - to name a few. While gamers and creators from around the world were amazed by the visual results of such characters, we raised a question- Did Keanu and Gina like their own digital doubles as they appeared in the respective games? This question is more relevant when the rapid advancements in computer graphics and artificial intelligence are dramatically transforming the accessibility of creating human digital doubles and how they are used, commonly known as avatars. No longer confined to the realm of specialized, high-cost equipment, these digital representations can now be generated using consumer-grade devices, making them more widely available to the general people. This democratization of digital double creation is not just revolutionizing entertainment industries, such as movies and video games, but is also fueling the burgeoning interest in the "metaverse"— a concept that envisions immersive, interconnected virtual worlds where individuals interact through their avatars ( Mystakidis 2022 ) . Given the current expansion of virtual avatar applications beyond video games, into various fields such as remote collaboration ( Pakanen et al. 2022 ; Higgins et al. 2021b ) , education ( Petrakou 2010 ; Hu et al. 2023 ) , virtual socialization ( Novak 2022 ) , and healthcare ( Fysh et al. 2021 ; Wonggom et al. 2019 ) , understanding user perception of these avatars is becoming increasingly important. As spatial computing and extended reality (XR) technologies continue to evolve, the potential applications of human digital doubles are broadening too, prompting a surge in not only academic endeavors but also in industry. Researchers are now focusing on both the psychological implications of these avatars as well as how users perceive them in various contexts. Previous avatar perception studies have primarily explored how varying levels of visual realism influence user perception, oftentimes using generic digital avatars that are not designed following the appearance of an existing human being ( Ferstl et al. 2021b ; Tinwell et al. 2013 ; Higgins et al. 2021a ; Amadou et al. 2023 ) . A number of studies investigated personalized digital avatars, digital doubles of existing people e.g. using photogrammetry but they usually fall short in terms of high-fidelity ( Higgins et al. 2021b ; Kim et al. 2023 ; Ma and Pan 2022 ) mainly due to access to high-end systems. Some studies investigated perception of stylized vs. realistic digital humans, such as- Fraser et al. ( Fraser et al. 2024 ) , whilst some investigated the perception of generic vs. personalized avatars, such as- Salagean et al. ( Salagean et al. 2023 ) . However, familiarity of the avatars is not investigated well. The extent to which an avatar resembles a familiar face— whether that of oneself, an acquaintance, or a stranger (analogous to generic avatars)— could significantly impact user perception. Although this familiarity aspect has been studied in the context of neural responses of participants for face recognition task of real human faces ( Campbell et al. 2020 ) , familiarity investigation in the context of avatar perception with realistic and stylized representations, remains underexplored. Insights gained from such research are vital for optimizing the use of human avatars in diverse applications related to education, remote collaboration, social XR, healthcare etc. In this study, we present a novel investigation into how avatar style (i.e., realistic vs. stylized) in conjunction with familiarity (i.e. self, acquaintance, unknown person) influence self/other identification, perceived realism, affinity, and social presence. Avatars are personal digital representations in a virtual 3D environment that can either have direct resemblance with a real-life person ( Salagean et al. 2023 ; Fraser et al. 2024 ) or can also appear as someone entirely different, oftentimes can be stylized characters or even animals ( Kyrlitsias and Michael-Grigoriou 2022 ) . In our study, we focus on the former- avatars having human resemblance with real-life, existing people. We conceptualize this category of human avatars as "Digital Doubles" as these specific human avatars are digital representations of the users having visual resemblance. In our study, we follow Fraser et al. ( Fraser et al. 2024 ) and use two versions of digital doubles in terms of appearance: stylized avatar (low human resemblance with high graphical resolution) and realistic avatar (high human resemblance with high graphical resolution). Unlike previous research that suggested a general preference for realistically rendered digital humans ( Pakanen et al. 2022 ; Higgins et al. 2021a ; Zibrek et al. 2019 ; Amadou et al. 2023 ) , our study specifically examines how these preferences shift when the digital double bears a familiar face. Our findings reveal that while higher appearance realism generally enhances self/other identification, perceived realism, and social presence, familiar avatars, especially those with high realism, paradoxically lead to lower levels of self/other-identification, perceived realism, and affinity. Interestingly, participants were consistently more critical of their own digital doubles, particularly when these had realistic style, but displayed greater tolerance and acceptance towards digital doubles of acquaintances or unknown individuals. 2. Related Work Previous related works pertaining to digital/virtual humans investigated both perceptual and psychological effects on users. Perceptual studies not only investigated the visual perception ( Zibrek et al. 2019 ; McDonnell et al. 2012 ; Ferstl et al. 2021a ) , but also auditory ( Mitchell et al. 2011 ; Suma et al. 2023 ; Maxim et al. 2023 ) . The other line of works looked into psychological aspects of digital humans including- trust ( Lin et al. 2023 ; Torre et al. 2019 ) , personality ( Pan et al. 2015 ; Thomas et al. 2022 ; Ruhland et al. 2015 ) , empathy ( Higgins et al. 2022 ; Higgins et al. 2023 ) and human

Read the complete text