Artificial Intelligence In Tariff Classification

Abstract

The current technological revolution and the rise of
artificial intelligence (AI) as a support tool in various
areas of knowledge and daily life force and encourage
companies to adapt to it as it is of relevance to promote
innovation and contribute to efficiency. The
implementation of AI in international trade seems to
have little opening since it is mainly limited to logistic
areas, while in other areas of study, such as tariff
classification, it has just been incorporated. The present
research work aims to experiment on AI tools,
specifically, as it is a public tool to be used, with the
objective of determining if this is an efficient tool for tariff
classification in terms of time, precision, accuracy and objectivity of the coding. Through various measurement methods such as the Likert scale and under a
comparison of codifications of different levels of
complexity obtained under the deductive method and
by GPT chat, the amount of information needed to
obtain an accurate codification was determined, and
thus it was concluded that GPT chat is efficient only in
terms of time. However, the constant use of the same,
driven by the branches of AI, supports that its
knowledge in this area is increased as more information
is provided, a fact that will favor the use of the tool in
order to obtain an accurate codification.