IIT Indore Develops AI Tool to Correct Gender Bias in Speech Translation

Researchers at the Indian Institute of Technology (IIT) Indore have developed an artificial intelligence tool called HGAST (Hierarchical Gender Arbitration for Speech Translation) to address gender bias in multilingual translation systems. The tool corrects errors that occur when English speech is translated into Indian languages such as Hindi, Marathi and Gujarati, where verb forms and sentence structures change based on the speaker's gender.

Unlike English, Indian languages use gender-specific verb forms. Many AI translation models are trained on datasets that favour masculine language patterns, causing systems to incorrectly generate male verb forms when translating women's speech. For example, in Hindi the verb "went" differs based on gender: "गया" (gaya) for males and "गई" (gayi) for females.

The research team, led by Dr Chandresh Kumar Maurya of IIT Indore along with collaborators, developed HGAST as a correction layer for existing translation systems. The tool works by detecting speaker gender through voice analysis and examining sentence structure to identify and apply correct gender-specific verb forms.