Google's image search algorithm displayed predominantly mugshots when searching for 'three black teenagers' while showing stock photos of smiling teens for 'three white teenagers', highlighting algorithmic bias in search results.
In June 2016, 18-year-old Kabir Alli from Virginia posted a viral Twitter video demonstrating that Google Image searches for 'three black teenagers' yielded primarily police mugshots, while searches for 'three white teenagers' returned stock photos of happy, smiling young people. The video was retweeted over 60,000 times and sparked widespread debate about algorithmic bias. Google responded that their search results reflect content frequency and descriptions across the web, stating that 'sometimes unpleasant portrayals of sensitive subject matter online can affect what image search results appear for a given query.' The company emphasized that results don't reflect Google's opinions or beliefs. Similar biases had been previously documented, including searches for 'unprofessional hairstyles for work' showing primarily black women and 'professional hairstyles' showing white women. Experts explained that the disparities result from search engine optimization, news coverage patterns, and the availability of stock photography, with black people comprising 13% of the US population leading to fewer stock images. The incident highlighted how algorithmic systems can perpetuate societal biases even without intentional discrimination.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Unequal treatment of individuals or groups by AI, often based on race, gender, or other sensitive characteristics, resulting in unfair outcomes and unfair representation of those groups.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed