abstract

The previous attempts of developing ethical accreditation of data have been focused on filtering out immoral banned words, which involves the difficulty of implementation in terms of effectivity and accuracy. This study aims to present the “Immoral sentence identification ontology” as a standard for filtering out immoral content including biased content in the process of classifying and organizing data using artificial intelligence technology. The immoral sentence identification ontology was designed to enable more systematic and accurate classification based on ethical theories by introducing various standards of moral judgment. We expect that this ontology will provide an effective standard for measurement of immorality in various everyday conversations and contribute to the establishment of an ecosystem where we can evaluate the degree of immorality of biased and unethical contents under broader circumstances.

keyword: ethical artificial intelligence, ethical accreditation, moral ontology, banned word, immorality