Home / Technology / Hindi AI Answers Lag English in Accuracy

Hindi AI Answers Lag English in Accuracy

Summary

  • AI chatbots performed worst in Hindi, with double error volume.
  • Chatbots struggled to find relevant information for Hindi queries.
  • Brands face reputational risk due to AI's language-specific blind spots.

A recent Stanford study indicates that AI chatbots exhibit lower accuracy when providing answers in Hindi versus English. Across six commercial chatbots tested, Hindi responses averaged 79.3 percent accuracy, nearly ten percentage points lower than other language regions, with Hindi performing worst overall. The primary issues identified were retrieval failures and source divergence, where chatbots could not find relevant information or found conflicting facts. English sources, such as Wikipedia, were predominantly cited for Hindi queries. This suggests AI models may default to English resources even when responding in Hindi, creating a structural flaw. For brands, this translates to inconsistent online visibility and reputation across different languages. Experts advise brands to build credible Hindi content beyond their own websites and ensure AI tools can access it, rather than treating it as a simple translation task. The findings underscore the importance of verifying crucial information, especially when seeking it in Hindi, as AI accuracy can vary significantly between languages.

Disclaimer: This story has been auto-aggregated and auto-summarised by a computer program. This story has not been edited or created by the Feedzop team.

Read more news on

Property Code: 5571