In this regard, he explained how the data which form the basis for algorithms reflect biases or systemic inequalities in the criminal justice system and it could lead to perpetuating the same bias and end up targeting neighbourhoods of marginalised communities.
“If historical crime data used to train these algorithms reflect biases or systemic inequalities in the criminal justice system, the algorithms may perpetuate these biases by targeting the same neighbourhoods as “high-risk” areas for future crime. This can result in disproportionate surveillance and policing of already marginalized communities, exacerbating social inequalities and perpetuating cycles of discrimination,” said the CJI.
Predictive policing algorithms often operate as black boxes, meaning their internal workings are not transparent, he added.
CJI Chandrachud was delivering the Keynote Address at the 11th Annual Conference of the Berkeley Centre for Comparative Equality and Antidiscrimination Law o the topic “Is there Hope for Equality Law?”
The conference was organised by the National Law School of India University, Bengaluru.
Source: Barandbench
