Research Project

The pervasive integration of artificial intelligence (AI) in high-stakes decision-making across the criminal justice, healthcare, and finance sectors has increasingly revealed how these systems can perpetuate and amplify existing societal biases. People of color are subjected to disproportionate surveillance and sentencing recommendations in the criminal justice system because predictive policing algorithms and risk assessment tools rely on historical crime data that is frequently biased by decades of over policing in marginalized neighborhoods. Similar to this, AI-based diagnostic tools and treatment recommendation systems that are trained on data that lacks racial and gender diversity may result in incorrect diagnoses, unfair treatment, and the perpetuation of health outcome inequities in the healthcare industry. Low-income and minority applicants are often disadvantaged by the banking industry's algorithms for credit score and loan approval, which frequently incorporate discriminatory patterns ingrained in economic histories. To address and prevent algorithmic prejudice, this paper examines shortcomings in the current legal frameworks governing artificial intelligence. Existing regulations frequently lack clarity, enforceability, and flexibility regarding the particular issues provided by machine learning systems, despite the expanding use of AI in delicate fields. This research emphasizes how jurisdictional discrepancies, ambiguous ethical standards, and antiquated regulatory frameworks all contribute to a lack of effective monitoring, which permits biased AI systems to function with little responsibility.