Chennai, 15 Oct 2025

IIT Madras researchers found bias in AI models used for legal tasks in India. They created a new Legal Safety Score to check fairness and accuracy of AI in Indian legal applications.

AI Shows Different Outputs for Similar Cases

The study tested AI models with two similar robbery cases. One case had a Christian name, the other a Hindu name. The AI said the robbery law applied to the Christian case but not the Hindu case. This shows AI can give unfair results based on a person's identity. The researchers say using such AI in real courts could lead to wrong decisions.

New Way to Make AI Safer for Indian Law

The team made a Legal Safety Score to check if AI is ready for Indian legal use. This score looks at how fair and correct the AI is. They also made the AI better by training it on Indian laws and situations. The researchers shared their work so others can use it too. They want people to know that AI trained on Western data may not work well for Indian laws without proper testing. The View From India Looking at World Affairs from the Indian perspective. First Day First Show News and reviews from the world of cinema and streaming. Today's Cache Your download of the top 5 technology stories of the day. Science For All The weekly newsletter from science writers…

Many had directly started using popular LLMs for legal purposes in India, even in court. Our work focused specifically on evaluating the readiness of LLMs for the Indian legal landscape, considering the unique diversity of India (caste, religion, etc.), rather than western scenarios.

Follow The Rising News for more verified updates on AI use in Indian legal systems. Follow The Rising News for more verified updates.

Updated 15 Oct 2025, 19:13 IST; source: link