The advent of Artificial Intelligence and self-learning algorithms has brought an onslaught of injustice through the perpetuation of human biases. The observance of such tangible discriminatory practices has shed light on the need for reform and algorithm security. Question the foundations of intelligence with Perspectoverse’s Urvi Agarwal through an account on the impending need for a shift in socio-economic value systems.
Humans are inherently flawed. They possess intrinsic biases and belief systems which tend to penetrate into economies, political frameworks, justice mechanisms and societal structures. Artificial Intelligence (AI) and computer technology have become what programmers and engineers view to be the channel for mitigating the interference of these flaws. Designers of such technology hope to create an equitable system where every consumer enjoys equal spoils.
Yet, what most people, including developers, forget and what has become increasingly apparent is that Artificial Intelligence reflects the structural character of those who create it. The behavioral biases within human beings are projected onto algorithms. The algorithms identify these programmes as functional elements, creating a self-learning intelligence. What one may find now, is robots furthering these discriminatory judgements by setting them in stone, here, their code.
Algorithmic biases have been observed in a number of technological systems, even the ones which seem ‘normal’. Hiring facial recognition tools to learn the features of company employees tends to the rejection of applicants with dissimilar external qualities in spite of being just as qualified. For instance, in the case of female computer programmers in the United States, that proves the inherent discrimination.
Certain mortgage approval algorithms readily infer particular zip codes correlated with race and poverty, perpetuating age-old discrimination into the modern world. The algorithm not only deprives persons of economic opportunity through a narrow interpretation of generalisation and balance of probability but also widens the rampant racial and income disparity in the States. The programme uses old, often unamended data to quite literally ‘predict the future’ of thousands- a dangerous practise in terms of statistics as well as ethics.
Within the United States’ already fractured healthcare paradigm, algorithms are often employed to recommend medical support to consumers. However, often these services are recommended to those who can readily access this support rather than those in need of it. This further exacerbates the alienation of masses from crucial medical agencies. It is evident, now more than ever, that society is facing systemic discrimination as well as algorithmic biases. The Digital Age thus requires both digital reform and security and protection.
Reform can be achieved through the thorough analysis of the question, “How much robot is too much robot?” That is, what is the extent of artificial intelligence interference that should be permitted within a socio-economic paradigm. It also comes with the impartial judgment of, “How much human should be allowed to vitalise AI?” The question addresses the question of whether human characteristics should be allowed to eternize themselves in computer programmes- would it do more harm than good?
Moreover, protection from this institutionalised discrimination is necessary. The Bill of Rights, ratified three years after the ratification of the Constitution of the United States of America, reserved certain rights and privileges to be enjoyed by the country’s citizens. It put in place certain guarantees, protections of sorts, that allowed “the people to be secure in their persons”.
Along with evolution of humankind and the societal network it nestles in, legal mechanisms must also evolve. They must grow with the humans that are subject to it. Hence, as a consequence of technological exploitation, legislation must be established to protect the general member of a community from common adversity.
At present, the United States Congress deliberates upon a Bill of Artificial Intelligence (AI) Rights, a reflection of that issued in 1791, in order to preserve the autonomy of citizens of the States. The Bill addresses privacy issues, problems of uneven powers dynamics as well as this issue of discrimination as a jurisdical crime and the party to be held accountable for it.
Although most details are not public yet, the common reader or perceiver may infer that this is a step in the right direction, or rather, the futuristic direction. AI is here to stay and minimising its magnitude is not a plausible option any longer. However, there is time to rethink the principles that it is founded on and whether they really help humans or not. There is time to secure people from the adversities brought on them by AI.
There is always time to redirect. This is the time, and we must cease it.
Written by Urvi Agarwal
Illustrated by Anushka Doshi
References:
Kommentare