Artificial Intelligence and Autonomous systems are distinct
Sagramsingh 19 [Raine Sagramsingh is a recent graduate of the FAMU-FSU College of Engineering. She graduated summa cum laude with a Bachelor of Science in mechanical engineering. She was an active member of the Florida State University Center for Leadership and Social Change’s Service Scholar Program. Raine completed internships with the Air Force Research Lab and Northrop Grumman Aerospace Systems.https://wise-intern.org/wp-content/uploads/2019/04/Raine_S_-FinalPaper.pdf]//LP
The terms in which the international community discusses lethal autonomous weapons systems (LAWS) can at times be ambiguous as there is an absence of a common vocabulary and common understanding of the terms autonomy and artificial intelligence (AI). LAWS are weapons systems that apply lethal force independently, with freedom from human control. AI has many subfields including machine learning, data mining, speech recognition, and image processing. Policymakers frequently refer to any of the subfields of AI as AI. By grouping these various terms together under the umbrella of AI, the term is frequently misused by technical experts, policymakers, and the general public. As advances in technology lead to increased autonomy in systems, the terms AI and autonomy are frequently used together. Despite their relationship to each other, it is a mistake to use these terms interchangeably. Currently, the international discussion does not define the technical characteristics of LAWS, hindering the development of public policies, laws, and practices related to the use of autonomy and AI in weapons systems. To improve understanding of the technical characteristics of autonomy and AI in LAWS, this paper aims provides the following recommendations. Recommendation 1: Classify the level of autonomy and intelligence in current weapon systems. Members of the international discussion on LAWS should evaluate where current weapon systems fall on the spectrum of intelligence and autonomy and should consider which of those should or should not be a part of the continuing discussion. Recommendation 2: Limit the use of the term Artificial Intelligence. While potential exists for fully-autonomous intelligent weapon systems, they are not a current reality. Using the term AI to refer to machine learning, computer vision, natural language processing, and many other fields of study limits the development of public policy because there is no individual evaluation for their potential positive or negative impact on LAWS. Recommendation 4: Distinguish autonomous systems from autonomous intelligent agents. An autonomous system does not imply the use of AI, whereas autonomous intelligent agent does. These terms represent significantly different technologies that warrant separate discussion. Recommendation 5: Distinguish between semi- and fully-autonomous systems. To make progress developing public policies regarding LAWS, the discussion must distinguish semi-autonomous systems from fully-autonomous systems. Regulating the two technologies with the same set of policies ultimately will be unsuccessful.