To examine reach and engagement performance among human-like female virtual profiles (VPs) on Instagram using a contemporary secondary dataset, emphasizing the significance of this dataset in understanding current trends.
Key Findings:
Photorealism and CGI-likeness co-occur with distinct reach and engagement patterns, indicating varying audience preferences.
Stable physical, ethnic, and behavioral presentation is frequently observed among high-reach and high-engagement profiles, suggesting consistency is key.
Posts featuring copresence, whether human-VP or VP-VP, are frequently observed among high-performing posts, highlighting the importance of interaction.
Specific appearance patterns, such as dark hair, dark eyes, and exaggerated body proportions, recur among top-performing VPs, indicating trends in audience attraction.
Interpretation:
The study aims to clarify which VP content deployment strategies and design signals are most prominent among high-performing accounts on Instagram, directly linking these strategies to the findings.
Limitations:
Previous literature often evaluated small samples or aggregated diverse character styles, limiting generalizability.
Mixed results in prior research regarding the effectiveness of VIs compared to human influencers, with specific examples illustrating these discrepancies.
Conclusion:
The findings contribute to understanding human-AI interaction and the emotional and behavioral engagement with AI-generated personas, while also suggesting potential avenues for future research.
