Physical characteristics of digital characters influence group categorization and recognition of affective states
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Ethnic bias in social group categorization and recognition of affective states persist in diverse countries like Canada, potentially affecting interactions with minority groups. With the growing use of digital characters (DCs) across various settings, it becomes crucial to explore whether these biases extend to virtual environments to mitigate these issues. This study created and validated 16 realistic DCs to examine how individuals perceive their physical characteristics while investigating the effects of ethnic biases. 112 participants from the majority group (White) completed a two-part online task in which they were asked to perceive in the 16 DCs 1) physical attributes in a neutral state such as phenotype (Black, White, Latin American, or Asian), gender, age, and realism, and 2) four affective states expressed by DCs (pain, anger, sadness, or neutral), as well as components associated with them (intensity, valence, and arousal). Participants categorized White DCs more accurately than Asian and Latin American DCs, and faster than Latin American DCs. The latter were also categorized less accurately and slower than the two other minority groups (Asian and Black DCs). Furthermore, the anger facial expression on Asian DCs was the least recognized among all other affective states and phenotypic groups. Thus, an attenuated own-phenotype bias emerged in contexts with multiple phenotypes, where very similar or very different physical characteristics contribute to efficient categorization. This study contributes to a finer understanding of how different phenotypic groups are perceived in virtual environments and introduces newly created digital characters that could be used for studies in human-agent interactions. • White digital characters are better categorized than those of minority groups. • Latin American digital characters had the lowest categorization accuracy. • Anger expressed by Asian digital characters is less recognized than other groups.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it