Artificial Intelligence as a Challenge to Social Justice in the Light of the Theories of J. Rawls and M. Walzer

• Název práce v češtině: Umělá inteligence jako výzva sociální spravedlnosti ve světle teorií J. Rawlse a M. Walzera
• Název v anglickém jazyce: Artificial Intelligence as a Challenge to Social Justice in the Light of the Theories of J. Rawls and M. Walzer
• Klíčová slova: Umělá inteligence, sociální spravedlnost, John Rawls, Michael Walzer
• Klíčová slova anglicky: Artificial Intelligence, Social Justice, John Rawls, Michael Walzer
• Akademický rok vypsání: 2023/2024
• Typ práce: bakalářská práce
• Jazyk práce: angličtina
• Ústav: Katedra politologie (23-KP)
• Vedoucí / školitel: Janusz Salamon, Ph.D.
• Řešitel: skrytý - zadáno vedoucím/školitelem
• Datum přihlášení: 30.04.2024
• Datum zadání: 30.04.2024
• Datum a čas obhajoby: 05.09.2024 07:00
• Místo konání obhajoby: C520, 520, seminární místnost IPS
• Datum odevzdání elektronické podoby: 30.07.2024
• Datum proběhlé obhajoby: 05.09.2024
• Oponenti: Ing. Petr Špecián, Ph.D.

Zásady pro vypracování

Topic Characteristics / Research Question:
Since matters of justice have been of concern to philosophers since the start of the evolution of the field, various theories of justice have been invented. While all those theories aim to address the contemporary injustices that prevailed at the time of their invention, they all have in common some basic assumptions such as the equipment of human beings with a certain set of unalterable natural abilities and talents that can either be cultivated or dismissed. However, some of those assumptions are likely to be challenged by the developments of Artificial Intelligence (AI). Consequently, the applicability of the existing theories of justice to the world we live in can be expected to become increasingly challenged as AI continues to be developed further.
In the first part of my thesis, I aim to explore how AI is likely to magnify social injustices and to identify which social goods, with a focus on the political and the economic sphere, are particularly vulnerable to the mentioned challenges of AI. Naturally, this is not only relevant to identify the goods that are to be protected more carefully but also serves as the point of departure for an investigation of how the existing theories of justice are limited in their conceptual framework to accommodate those challenges.
Following this, in the second part, I will work with John Rawls’ theory of ‘justice as fairness’ as well as Michael Walzer’s ‘justice as complex equality’, two of the most influential proponents of the academic debate between liberals and communitarians on justice and show how both approaches to matters of justice are conceptually insufficient to accommodate the threats AI poses to social justice from a more theoretical perspective.
This will be valuable since the political solution to matters of social justice always needs to be grounded in some theoretical framework that guides how to reach optimal outcomes for the individual/society. However, human beings are social animals that cannot flourish completely isolated from another so questions of social justice are, and will remain, at the center of human life. A firm and robust theory of justice is needed to ensure human flourishing.
Therefore, to provide a base for the invention of such a theory, the aim of this thesis is to answer the following research question: What are the conceptual limitations of the existing theories of justice in light of AI developments and will it be enough to amend those theories, or do they need to be replaced by an entirely new theory of justice?
However, to answer the bigger research question, various other questions need to be answered in the process. The just distribution of which social goods are most vulnerable to developments of AI? In what way is AI threatening them? Which of the existing theories of justice is most promising to be able to keep up with the challenges?
It is worth mentioning that the goal of this thesis is by no means to address all existing theories of justice nor to invent a new one or to propose how Rawls’ and Walzer’s theories should be amended but only to identify the conceptual limitations thereof. Furthermore, it is built on the assumption that AI will continue to develop the way it is currently being done and does not assume the introduction of a legal and ethical framework (additional to the existing one) to regulate the development of AI.

Working hypotheses:
1. Existing theories of justice cannot be applied to injustices arising from automated decision-making, bias, and a shift in accountability/responsibility caused by developments in AI.
2. AI developments challenge the very foundational assumptions the assessed theories of justice are built on. Those include equal access to information, and equality of opportunity as well as given, unalterable individual talents that cannot be changed but only cultivated or dismissed.
3. It is not enough to only amend the existing theories of justice but, considering the developments of AI, new theories of justice that are flexible enough to be adapted to volatile (social) environments need to be invented.

Methodology:
In the first part, I will follow M. Walzer’s methodology of dividing (public) life into various spheres of justice. In my thesis, I will focus on the economic sphere as well as the political one and identify how unjust social practices are being amplified within them through the current developments of AI. This will be examined through the lens of the Rawlsian method of reflective equilibrium.
Following from part one, I will assess Rawls’ and Walzer’s theories by employing their own methodologies. I will investigate whether Rawls’ theory can adequately respond to the challenges AI poses on social justice by asking whether the challenge of AI challenges the plausibility of Rawlsian theory according to his own method of moral investigation, the reflective equilibrium. The same will be done for Walzer: I will investigate how the spheres of justice will be linked through the developments of AI by sticking to his method of deep interpretation.
In the last part, I will provide a comparative analysis of both theories and examine which of them is conceptually less limited to address the challenge of AI.

Seznam odborné literatury