Event / Technology

Google AI Bias Safety Issue

← All events

Updated

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.

0 reported locations
Geography of the reported facts

No reported coordinates available.

The story so far

Daily activity

1 articles
0 articles

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.

What happened

The facts
  1. Google's AI updates responses to remove ethnic and national discrimination

    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.

    1 corroborating report

  2. Google acknowledges AI search results 'not up to standard' and commits to fixes

    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.

    1 corroborating report

  3. Google's AI generates biased safety warnings based on nationality and ethnicity

    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.

    1 corroborating report

Read the coverage