The Illusion Of Neutrality: How ‘Objective’ Algorithms Perpetuate Discrimination
Dishi, BBA LLB, Symbiosis Law School, Pune
Introduction
In 1936, the mathematician Alan Turing specified algorithms as determinate sequences of computational steps, mechanical procedures that progress toward logical conclusions with mathematical exactitude, insulated from subjective human judgment. This foundational characterization positioned algorithms as instruments of pure rationality, detached from the social contexts in which they would later operate. For decades, Western technological discourse sustained this view: algorithms were cast as objective arbiters capable of transcending biases embedded in human decision-making systems. When automated risk assessment entered criminal justice in the 1990s, when financial institutions mechanized credit- worthiness determinations in the 2000s, and when law enforcement agencies implemented predictive policing architectures in the 2010s, the underlying premise remained ever constant to imply that algorithmic systems would achieve fairness through their non-human nature.
India’s relationship with this technological narrative has been peripheral and central at the same time. While the Global North was busy building out algorithmic infrastructure, India became its key engineering workforce, with the IT services sector alone employing more than five million professionals by 2020. But this technical capability has not been matched by a similar critical inquiry into algorithmic neutrality in Indian contexts. When Aadhaar’s biometric authentication system started deciding who got food subsidy, when automated loan appraisal platforms began considering creditworthiness for first-time borrowers from marginalized communities, and when algorithmic hiring tools entered the recruitment process of Indian corporations, notions of objectivity were imported lock, stock, and barrel without interrogation of their founding premises.
This essay argues that discrimination in algorithms is sustained not through technical failure but through structural design-through the ossification of historical inequities within training data, through the deployment of socioeconomic proxies that reproduce caste and class hierarchies, and through operation within legal and technical opacity that precludes accountability. The Indian context, analysed later in this essay, manifests how algorithmic bias intersects with pre-existing systems of social stratification to magnify rather than mitigate discrimination.
