|
|
The transformative capacity of artificial intelligence: A critical approach to alignment in Security Studies
Morales Mendoza, Ivan Emmanuel ; Střítecký, Vít (advisor) ; Judge, Andrew (referee)
The Transformative Capacity of Artificial Intelligence: A Critical Approach to Alignment in Security Studies. By Ivan Emmanuel Morales Mendoza Charles University Student Number: 81738503 Abstract As the capabilities of Artificial Intelligence (AI) systems increase around the world, academics and policymakers have paid renewed attention to ensure their behaviour remains aligned to their operator's expectations, especially in sensitive areas involving public, national, and international security. The use of AI systems in security contexts, especially those considered as "high-risk," is a security issue because humans are trusting a technology, based on their expectations, to produce outcomes to make individuals and communities safer. This leads to an increase in the use of algorithms to perform policing duties, target identification, intelligence generation, among others. However, current theoretical and methodological proposals are limited when considering the transformative impact AI can have in security studies. Therefore, this article will analyse three of the current proposals for AI alignment -symbolic, Machine Learning and hybrid- to illustrate how security studies could strengthen its conceptual and methodological robustness when analysing security risks associated with the use of AI systems. To do...
|
guest ::
