Image Cropping on Twitter: Fairness Metrics, their Limitations, and the Importance of Representation, Design, and Agency

Abstract

Twitter uses machine learning to crop images, where crops are centered around the part predicted to be the most salient. In fall 2020, Twitter users raised concerns that the automated image cropping system on Twitter favored light-skinned over dark-skinned individuals, as well as concerns that the system favored cropping woman's bodies instead of their heads. In order to address these concerns, we conduct an extensive analysis using formalized group fairness metrics. We find systematic disparities in cropping and identify contributing factors, including the fact that the cropping based on the single most salient point can amplify the disparities because of an effect we term argmax bias. However, we demonstrate that formalized fairness metrics and quantitative analysis on their own are insufficient for capturing the risk of representational harm in automatic cropping. We suggest the removal of saliency-based cropping in favor of a solution that better preserves user agency. For developing a new solution that sufficiently address concerns related to representational harm, our critique motivates a combination of quantitative and qualitative methods that include human-centered design.

Accepted to CSCW 2021

References in corpus (1)

• Towards a Critical Race Methodology in Algorithmic Fairness

Cited by in corpus (12)

• Who Audits the Auditors? Recommendations from a field scan of the algorithmic auditing ecosystem
• Speciesist bias in AI -- How AI applications perpetuate discrimination and unfair outcomes against animals
• Cruising Queer HCI on the DL: A Literature Review of LGBTQ+ People in HCI
• Algorithmic Auditing and Social Justice: Lessons from the History of Audit Studies
• How Platform-User Power Relations Shape Algorithmic Accountability: A Case Study of Instant Loan Platforms and Financially Stressed Users in India
• Participation and Division of Labor in User-Driven Algorithm Audits: How Do Everyday Users Work together to Surface Algorithmic Harms?
• Beyond Behaviorist Representational Harms: A Plan for Measurement and Mitigation
• Wikibench: Community-Driven Data Curation for AI Evaluation on Wikipedia
• Having your Privacy Cake and Eating it Too: Platform-supported Auditing of Social Media Algorithms for Public Interest
• The Fall of an Algorithm: Characterizing the Dynamics Toward Abandonment
• Discrimination through Image Selection by Job Advertisers on Facebook
• Causal Inference Struggles with Agency on Online Platforms