Large Language Models Show Poor Accuracy in Cosmetic Chemistry Advice

A new study published on arXiv evaluated the accuracy of 14 large language models (LLMs) in providing information on cosmetic chemistry and skin health. The research found that these general-purpose LLMs performed poorly, particularly in quantitative reasoning and identifying chemical structures. While they could answer general skincare questions reasonably, their responses lacked the technical depth necessary for consumers to make informed decisions. The study suggests that LLMs are not currently reliable sources for cosmetic chemistry information and highlights the risk of authoritative-sounding but technically inaccurate outputs. AI

IMPACT General-purpose LLMs are currently unreliable for specialized technical advice like cosmetic chemistry, posing risks due to authoritative-sounding inaccuracies.

RANK_REASON Academic paper evaluating LLM performance on a specific domain. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Ai Chatbots
• arXiv
• cosmetic chemistry
• large-language models
• skin health

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →