• Kshitij Bhaskar
• Kaleb Waugh
• Mahendra Lembhe
• Hariom Yeslote
• H.R.Kulkarni,Madhuri Pise

Text datasets form the foundation of Natural Language Processing (NLP), powering applications from online search and recommendation engines to decision-critical systems in healthcare, law, and finance. However, these datasets often encode social, cultural, historical, and annotation-driven biases that, when learned by AI, propagate and amplify unfair outcomes. This paper presents a comprehensive and comparative review of bias detection and mitigation in text datasets. It surveys major forms of bias—representational, stereotypical, sampling, annotation, cultural, and algorithmic—while analyzing state-of-the-art detection tools such as statistical audits, embedding association tests (e.g., WEAT, SEAT), explainable AI (XAI), and behavioral probing. It further examines mitigation strategies across the NLP pipeline: pre-processing (data balancing, counterfactual augmentation), in-processing (adversarial training, fairness-aware objectives), and post-processing (threshold calibration, output rewriting). The review critiques benchmark datasets like StereoSet, CrowS-Pairs, Jigsaw Toxicity, BEADS, and domain-specific corpora. Persistent gaps include intersectionality, multilingual fairness, dataset documentation, annotation bias, and the need for dynamic, real-time monitoring in deployed systems. The paper concludes with future research directions emphasizing living benchmarks, participatory AI, explainable fairness, and continuous bias auditing.

Bias Detection And Mitigation In Text Datasets Using Natural Language Processing: Comprehensive And Comparative Review

"Bias Detection And Mitigation In Text Datasets Using Natural Language Processing: Comprehensive And Comparative Review", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 1, page no.146-156, January-2026, Available :http://www.jetir.org/papers/JETIRHG06015.pdf

"Bias Detection And Mitigation In Text Datasets Using Natural Language Processing: Comprehensive And Comparative Review", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 1, page no. pp146-156, January-2026, Available at : http://www.jetir.org/papers/JETIRHG06015.pdf