Member-only story
From Models to Realities: Ethical Challenges in the Application of Deep Learning in Politics, Law and Society.
In recent years, the application of deep learning models has permeated a wide array of seemingly unrelated fields, from medical diagnosis1 and improved algorithms for solving linear algebra problems2, to everyday applications such as ChatGPT assisting in trip planning based on our context.
This document explores the applicability of deep learning in the legal and political sphere, considering its potential utility in the domains of laws and political order. These fields, like any systematically organized field belonging to a closed knowledge community (comprising lawyers and political scientists), possess a strong esoteric component, particularly in language, potentially creating a barrier to justice for those not versed in these fields but interested in exercising their rights.
It also opens the door to automating biased legal processes, potentially leading to increased segregation of vulnerable populations and amplifying the bias these models already have in the population, such as the classification of dangerous zones in cities.
TOOLS OF JUSTICE OR TOOLS OF SEGREGATION?
Models, regardless of their nature, are partial representations of reality; we cannot…
