Factors Influencing Data Partiality in Artificial Intelligence

This study proposes a conceptual framework to investigate factors influencing the data partiality in Artificial Intelligence (AI). However, the academic research on data partiality focusing on AI is limited across the bibliographic database sources. This study aims to address the gaps by proposing a developed framework that integrates three factors: the AI algorithm, black data, and user revise terminology highlighted in the past literature. The AI algorithm refers to the issues on the training data as a dataset used in the tools, which stimulates the data partiality as the outcome retrieved by the user. The black data is influencing data partiality on the existence of unknown data. The user revise terminology represented on the keywords used by the user to search for information, which incorrect keywords with not specify will lead to the AI to give all related information as an output without filter. The framework asserts that these three elements directly affect the partiality of data in AI. A quantitative methodology will be used in this study to cover the collection of survey data from the community under the MDEC program called Global Online Workforce (GLOW). The framework contributes a theoretical understanding of AI algorithms, black data, and user-revised terminology that influence data partiality in AI. In future research, the framework can be extended to test the data partiality in AI tools used in information agencies, as these bodies govern the safeguards of the accuracy of the information.