Google's AI System Generates Biased Safety Warnings by Nationality
Google found its AI generated nationality-based bias in safety warnings and committed to fixes.
Event / Technology
Google's AI system generated ethnically and nationally discriminatory safety warnings, with systematic bias favoring certain groups while generating police warnings for Africans and Middle Easterners. Google acknowledged the issue and committed to fixes.
No reported coordinates available.
2026-08-26 to 2026-11-08 · UTC publication dates
Coverage begins 26 Aug, 15:05 UTC
Follow the documented timeline or explore the related reporting.
Following the controversy, Google's AI system updated its responses for several previously stigmatized nationalities. The updated responses now state there is nothing particular to do and that a person's origin or nationality does not determine their behavior or kindness.
Google responded to the controversy by acknowledging that the AI search results were 'not up to standard' and stated the company is working to remedy the issue. The company stated on X that results can vary considerably from search to search and denied that the inconsistent warnings targeted any particular group.
Google's AI system produced inconsistent and discriminatory safety warnings when asked about interactions with people from specific nationalities. Testing by multiple news organizations, including Libération's CheckNews service, which tested approximately 60 different nationalities, found that the AI's responses varied significantly based on nationality, skin color, or religion mentioned in the query. Specifically, mentions of Malians, Ivorians, or Senegalese people systematically triggered warnings to call police and assess potential feelings of insecurity.
Google found its AI generated nationality-based bias in safety warnings and committed to fixes.