Sima AIunty: Caste Audit in LLM-Driven Matchmaking

1. 1Investigate how LLMs reproduce or disrupt caste-based stratification in matchmaking
2. 2Analyze the cultural and historical context of caste in South Asian marital decision-making
3. 3Develop and apply auditing methods to detect and mitigate caste bias in LLM-driven matchmaking
4. 4Evaluate the impact of LLM-driven matchmaking on social inequalities and cultural norms

AI engineers and researchers working on LLMs and social applications can benefit from understanding the potential biases in their models, while product managers and designers can apply these insights to create more inclusive matchmaking platforms

💡 LLMs can reproduce or disrupt caste-based stratification in matchmaking, highlighting the need for auditing and mitigation strategies

🚨 LLMs in matchmaking may perpetuate caste bias 🚨