GNAI Visual Synopsis: A collage of diverse faces with some images clearly altered – lighter skin, blue eyes, slimmer faces – contrasted with unedited, genuine portraits, showcasing a visible dichotomy that reflects the article’s theme of AI bias.
One-Sentence Summary
Tricia Crimmins from web_crawlr examines the racial and fatphobic biases in a new AI-based TikTok filter that alters user appearances. Read The Full Article
Key Points
- 1. Tricia Crimmins, contributing to web_crawlr, discusses “Problematic on TikTok,” focusing on an AI filter that has been used nearly 4 million times but demonstrates problematic biases.
- 2. The filter, intended to give a vintage photo appearance, has altered user features in ways that align with racist and fatphobic prejudices, such as lightening skin tones and slimming facial shapes.
- 3. TikTok user E.L. Shen revealed how the filter consistently changed her appearance to resemble a white woman, problematic given historical anti-Asian discrimination, while user Abby Morris highlighted the filter’s fatphobic tendency to thin out faces.
Key Insight
The pervasive nature of these biases in advanced tech like AI filters underscores a larger systemic issue of discrimination embedded within the algorithms, primarily inherited from historical and societal prejudices.
Why This Matters
The discussion is vital as it reveals a broader problem with AI technology mirroring societal biases, impacting marginalized communities and perpetuating discrimination. It calls for the urgent need to address these issues in technology development, fostering inclusion and fair representation in AI applications.
Notable Quote
“You will not convince me to get on board with AI until one of these trends keeps me fat,” – Abby Morris highlights the fatphobia embedded within the new TikTok AI filters, stressing the need for body inclusivity in technology.