Quantitative Measurement of Bias in AI-Generated Content: A Comprehensive Narrative Literature Review
Ashish K. Saxena · 2024
Corporations, institutions, and individuals increasingly use artificial intelligence (AI) to make decisions and predictions that shape many aspects of human lives. Furthermore, individuals and organizations use AI to generate articles, blog posts, social media posts, or books, which can be substantial in various scenarios. However, this AI-generated content is often subject to biases that have unintended consequences on human lives during decision-making. The ability of AI models and systems to sustain and augment biases is a growing issue. Therefore, this literature examines how bias in AI-generated content can impact society, the sources of bias, and the quantitative methods used for measuring the bias. The literature also reviews real-world scenarios where quantitative bias measures in AI-generated content have been implemented successfully.