Generative AI in Cosmetics Regulations: A Comparison Between ChatGPT, Bard, and Claude

Recent advances in generative AI systems tools ChatGPT, Claude, and Gemini/Bard have shown promising capabilities across industries, including potential applications in cosmetics for regulatory oversight and consumer education. This paper analyzes the competence of these systems in responding to queries related to regulations in the cosmetic industry. Fourteen FAQ questions on cosmetics regulation by the U.S. Food and Drug Administration (FDA) and 15 by the Indian Central Drugs Standard Control Organisation (CDSCO) were tested for Readability, Tone, Accuracy, and Response Style. Flesch- Kincaid Readability Test, New Dale-Chall Readability and Spache Readability Score were used to test these. The analysis suggests that Gemini/Bard scores high on Readability and Response Style, ChatGPT scores high on Accuracy, and Claude scores high on Tone.