Abstract

This chapter addresses the proliferation of AI generated avatars on social media and the tendency of generative systems to reproduce hypersexualized femininity . Adopting a two layer defaults framework the authors analyze how synthetic defaults stemming from model priors and algorithmic defaults driven by platform visibility regimes jointly incentivize eroticized portrayals that undermine credibility . To proffer a practical alternative this chapter documents the design of Alicia which is a counter default avatar created through a two round Delphi study with seventeen experts . The study translates consensus into operational constraints such as functional styling and adult age coding and implements a consistency pipeline to resist identity drift . By mapping these decisions to relational outcomes the authors argue that resisting hypersexualization is a necessary condition for ethical relationship building in synthetic environments.