Archive · 16 Sep 2025
AI in India: Caste Bias and the Need for Auditing
AI, the computers that help make decisions, are starting to show signs of prejudice in India.
This is because the AI is trained on data that contains biases.
If the data shows that certain groups of people are seen as lower than others, the AI will learn to repeat those biases.
This can cause problems for people, such as not getting jobs or loans.
To fix this, experts suggest making sure AI is checked for biases and the data used to train the AI is inclusive.
This ensures everyone has a fair chance.
AI systems in India are learning and reproducing caste-based discrimination.
Algorithms are trained on biased data reflecting existing societal biases.
Biases manifest in job applications, credit approvals, content recommendations, and image generation.
Lack of transparency and accountability exacerbates the problem.
Auditing AI for bias, diverse datasets, and Dalit representation are crucial for inclusivity.
- Who
- AI systems, companies, and society.
- What
- AI systems learning and reproducing caste-based discrimination.
- When
- Present and future implications of biased AI systems.
- Where
- India.
- Why
- Algorithms are trained on data reflecting existing societal biases.
- Issue
- AI systems in India exhibiting caste bias due to biased data.
- Impact
- Algorithms perpetuate discrimination in jobs, loans, and visibility.
- Examples of Bias
- Stereotypical image generation, content moderation, and algorithmic curation favoring mainstream content.
- Proposed Solution
- Auditing AI for bias, diverse datasets, transparency, and Dalit representation in tech.
- Digital Systems Affected
- ATS, Recommendation engines, ad delivery optimisers, fintech, social media, and AI image/video models