/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

Twitter will no longer crop photo previews on the web after a similar update on iOS and Android in May, following claims of bias in its auto-cropping algorithm

Emma Roth / The Verge :

The Verge Emma Roth

Context & Ripple Effects

Twitter's image-cropping pipeline began in 2018, when it adopted machine learning to identify the most salient part of a photo for preview thumbnails (machine-learning-based salience cropping). Criticism of racial bias in that system surfaced in October 2020, prompting Twitter to defend the approach while promising users more control.

The company then published an investigation in May 2021 finding the algorithm favored white people and women (the bias investigation results), and by March had begun testing full-size image display alongside 4K uploads on mobile (large-image testing on iOS and Android). Today's change extends that replacement to the web, completing the retirement of auto-cropping across platforms.

First-order effects

  • Web users now see uncropped photo previews by default, matching the behavior already shipped on iOS and Android in May — the salience-cropping model Twitter built in 2018 is no longer deciding what parts of images get shown.

Second-order effects

  • Other platforms running similar machine-learned preview or ranking features face the same audit-or-replace dilemma Twitter went through: publish bias findings and risk the headline, or quietly move control back to users.

Third-order effects

  • If the pattern holds, automated aesthetic judgments baked into consumer products become a liability class of their own — companies increasingly substitute deterministic, user-controlled presentation for learned 'salience' once an algorithm's failures are documented publicly.

The trend: Consumer platforms are retreating from machine-learned content presentation toward user-controlled defaults whenever algorithmic bias is demonstrated, with Twitter's crop-to-full-image migration as the template.