Prospects of Artificial Intelligence in the Annihilation of Caste

The Context

On the 127th birth anniversary of Baba Saheb Bhimrao Ambedkar, his supporters gathered at the United Nations Headquarters in New York to commemorate the day. The event, themed ‘Artificial Intelligence for Humanity’, aimed to highlight the potential of Artificial Intelligence (AI) in addressing social issues such as the caste system and reducing inequalities in India by 2030. The event was scheduled to be coordinated by the Foundation for Human Horizon-USA in collaboration with the United Nations Department of Economic and Social Affairs. The event was anticipated to be attended by AI experts and representatives from Microsoft, Google, Facebook and other such companies.

Deelip Mhaske of the Foundation for Human Horizon-USA said

AI, if used properly, can be the main driver to reduce inequalities as it can create and protect much-needed safe zone, i.e., the AI prediction of caste violence by zip code, or caste group can provide tools to develop preventive measures to stop caste violence similar to the western world where they are using “crime maps” or “diseases map”.

Furthermore, he also stated that ‘there is great research going on AI’s use in preventing racial discrimination by IBM’s “optimized pre-processing for discrimination”’. ‘Similarly, we can use AI to reduce caste biases and violence’, he said, adding that Google had developed a strategy for ‘equal opportunity by design’, an AI-based platform.

Noting that the AI diversity foundation has been very active in using AI to tackle discrimination in all aspects of life, Mhaske said the Foundation for Human Horizon was exploring a partnership with Google, IBM, Facebook and Microsoft to use AI to eradicate caste biases and discrimination (PTI, 2018).

The expectation to get away with the caste system is largely a hope to curb the discrimination faced by the members belonging to the vulnerable group.

Artificial Intelligence

‘The science of making machines do things that would require intelligence if done by men’ is what Marvin Minsky refers to as AI. In the field of computing, Minsky and his associates were outcasts. AI experts talked of replacing the human mind, a ‘meat machine’, with their more effective electronic models, producing nothing less than a new species for the globe, while others applied computing skills to business and engineering.

It should come as no surprise that experts in AI do not regard themselves as contemporary counterparts of the toy manufacturers who entertained European aristocracy in the bygone ages. Their work with computers undoubtedly calls for a level of education and intellectual rigour that toy-making does not. Compared to the early industrial revolution, our society is significantly more reliant on science and technology, and technologists enjoy a greater standing. AI programmers are especially worried about their status because even other computer specialists have questioned their work. They want the proper respect since they are toymakers who deal with the priciest, most powerful and most prestigious toys we call computers. They attempt to portray the study of thinking systems or the science of cognition as a young discipline in their books. They contend that no theory of memory, learning, language, or human inference can be more precisely and empirically tested than AI. They claim that a good programme is more than just a machine–human metaphor or suggested parallel. Instead, it is so similar to a man or woman in significant ways that we can learn about the human mind, or rather about cognitive processes in general, of which computers and people are both examples, by seeing how they perform (Bolter, 1984).

AI refers to the field of study and development focused on creating intelligent machines. Intelligence, in this context, is the ability of an entity to effectively and proactively operate within its surroundings. A common misconception about AI is that technology will allow computers to think like people. According to the well-known Turing test, AI is attained when a person is unable to distinguish between a machine and a human when a query is answered. Some people use the phrase to describe computers that process vast amounts of data, make deductions and gain expertise using algorithms.

Some people think AI is headed towards creating intelligent machines that will be considerably more powerful than humans. Following this technological singularity, computers will develop further and give rise to swift technical advancements that will drastically and unpredictably alter humankind. According to some analysts, the singularity might happen as early as 2030 (Etzioni & Etzioni, 2017).

Caste

In a closed system of social stratification, caste separates and deprives those who are at the bottom of the caste hierarchy by assigning employment based only on birth and enforcing a rigid code of behaviour. The hierarchy of the caste system is arranged inside the Varna system, which is the four-fold occupational order. Each Varna contains hundreds of jaati, or caste groupings. Despite not adhering to caste and being outside of its fold, this system marginalizes Adivasis (indigenous people), nomadic and semi-nomadic tribes. The caste system was welcomed by British colonial officials because it offered a useful framework for selecting subjects under their authority. Colonial administrations selectively policed populations whose behaviours they deemed to be illegal, threatening, or abnormal to create the appearance of order with the few resources at their disposal. The colonial State’s monopoly was threatened in particular by semi-nomadic and nomadic tribes, whose unrestrained, mobile economies allowed them to avoid paying state taxes. Thievery, dacoity and robbery were constructed by colonists as occupational crimes due to the strict hereditary occupational order of the caste system. Moreover, the colonial illusion of the dangerous and hereditary criminal disposition of nomadic and semi-nomadic tribes was influenced by their mobility and lack of caste capital (Sonavane & Bej, 2021).

The caste system has often been regarded as an outdated institution and a historical source of disadvantage. However, this perspective fails to acknowledge its ongoing significance as a system that provides advantages and promotes discrimination in the modern economy, particularly after the liberalization period starting in 1991. Interrogations arise regarding caste as a social stratification system, the impact of caste on rural inequalities after liberalization, its effect on urban labour markets and the business sector and the consequences of affirmative action laws on education and employment in the public sector. The institution of caste is a complicated phenomenon that is both undermined and revitalized by current economic and political factors. It plays a significant role in perpetuating socioeconomic and human capital inequities at the national level and has a profound impact on subjective welfare. Caste effects are not limited to specific locations; they extend from rural areas to urban areas and permeate practically all sectors of society. The persistence of caste in the era of the market can be attributed to its inherent benefits. The discriminatory practices associated with caste enable certain individuals to monopolize opportunities, while any progress made by marginalized groups is met with degrading violence. The situation indicates that policy innovation is necessary to tackle both market and non-market discrimination and to eradicate obstacles, particularly in the private and informal sectors. Additionally, it is crucial to guarantee that the issue of caste is appropriately addressed in the global development policy discourse.

Multiple global human rights organizations assert that more than 260 million individuals worldwide experience discrimination based on caste, also known as ‘work and descent’ according to the United Nations. Caste is considered a crucial factor in social exclusion and development, impacting approximately 20%–25% of the global population. This includes, but is not limited to, the people of South Asian nations and their diasporas (Mosse, 2018).

Presently, the so-called higher castes possess the majority of the nation’s capital assets, including land, buildings and financial resources. Conversely, the so-called lower castes primarily engage in the economy as wage labourers. As we descend the hierarchy, there is a decline in both per-capita income and the availability of high-status employment. Likewise, the advantages of factors such as improved education or capital holdings also diminish. Simultaneously, the percentage of those residing in poverty rises. This signifies the presence of a hierarchical structure of ‘graded inequality,’ as elucidated by Dr B.R. Ambedkar.

The social policy regarding caste, as well as the guidelines set by the concerned authority and stakeholders, primarily concentrate on the disadvantages faced by specific groups. They view caste as a fixed or remaining issue that can be resolved through remedial measures, protective measures, safeguards and the handling of complaints. However, they fail to recognize caste as a constantly changing issue that should be addressed to tackle inequality and discrimination in the economy and society (Mosse, 2018). Although the 1989 Prevention of Atrocities Act in criminal law forbids some actions against members of Scheduled Castes (SCs), there is no comprehensive legislation that specifically addresses caste-based prejudice and promotes equality. The daily disparities of caste are generally seen as issues that can be addressed by societal and, more specifically, market-driven changes (Mosse, 2018).

Technology improvements will inevitably have a negative impact on society if there are no well-organized and morally led approaches to the field. As the pinnacle of technical advancement, AI and machine learning could end up as instruments of power controlled by a small number of powerful individuals if they are not created according to non-discriminatory standards. Similar to how the core idea of capitalism is to concentrate resources and wealth in the hands of a small number of people, AI may take a similar course if it is not handled fairly and equally.

The biased data used to train AI systems is the cause of the racial bias that has surfaced in AI applications as a significant problem. Because biased data is being used in future AI systems, similar bias problems are probably going to arise in India. Experts and academics have advocated for anti-caste tech laws to end caste-based prejudice in emerging technologies. To guarantee justice, equity and equal opportunity for Dalits and other lower castes who experience institutional oppression because of their caste identity, caste-based discrimination in AI algorithms must be reduced to an absolute minimum. The problem of caste discrimination is especially significant in India, as lower castes and Dalits experience systemic marginalization and oppression (Kamble, 2023).

Future Perspective

The concerns about the elimination or breakdown of the caste system or the caste customs in daily life illustrate many ideas about social norms that are considered acceptable. Since AI is a technological instrument, it will either make the caste system more rigid and functional or turn it into a social threat in each situation where AI’s role in the caste is determined.

The role of Indian caste-based prejudice in perpetuating Silicon Valley’s diversity challenges is rarely addressed. The demographic composition of the technical workforce in the IT industry has been shaped by long-standing labour practices, including the recruitment of graduates from prestigious colleges and reliance on the H-1B visa programme to attract highly skilled professionals. Dalit engineers and their sympathizers argue that software corporations, despite being aware of caste bias, have not explicitly prohibited it. Nevertheless, the Dalit rights movement has recently broadened its scope, acquiring a worldwide perspective and demanding corporate America enact change.

Nearly 260 US tech professionals filed complaints about caste inequality with Equality Labs, a non-profit advocacy group for Dalit rights, three weeks after the lawsuit was publicized. These concerns were either through the company’s website or in emails to specific staff members. According to executive director Thenmozhi Soundararajan, allegations included sexual harassment, biased peer reviews, bullying, discriminatory hiring procedures and remarks and jokes based on caste. The highest number of claims, totalling 33, were submitted by employees of Facebook. This was followed by employees from Cisco with 24 claims, Google with 20 claims, Microsoft with 18 claims, IBM with 17 claims and Amazon with 14 claims. Every firm asserted that discrimination is unacceptable.

Additionally, in a statement shared exclusively with The Washington Post, a group of 30 Dalit female Indian engineers who work for Google, Apple, Microsoft, Cisco and other big corporations claims to have experienced caste bias in the US tech industry (Tiku, 2020).

The eradication of caste using AI entails harnessing it to tackle and eradicate prejudice and inequality based on caste. To mitigate caste biases and violence, one can employ AI technologies in various domains, including job and loan approval procedures, social media surveillance and the formulation of anti-caste ethical principles for emerging technology. Efforts such as ensuring fairness in algorithms, establishing ethical frameworks and applying Dr B. R. Ambedkar’s perspectives on social justice can help in building a digital India that is more inclusive and equitable. Furthermore, the implementation of anti-caste technological regulations in the metaverse, as well as the use of AI to eliminate caste prejudices and discrimination, are essential measures in the pursuit of eradicating caste through contemporary technology.

However, the companies mentioned above and other similar companies are directly or indirectly disseminating AI. How will eradicating casteism with the help of AI be fulfilled?