This paper explores the potential of Large Language Models (LLMs) to achieve fairness in classification tasks through in-context learning. The study investigates whether LLMs can incorporate fairness considerations into their responses when prompted with specific fairness rules. The research focuses on assessing the fairness of LLMs in classification tasks, particularly in the context of the UCI Adult Income Dataset, which contains demographic and financial information of individuals. The study compares the performance of three LLMs: GPT-4, LLaMA 2, and Gemini, in terms of accuracy and fairness metrics across zero-shot and few-shot learning setups. The results indicate that GPT-4 performs the best in terms of both accuracy and fairness, while LLaMA 2 and Gemini show suboptimal performance. The study also introduces a framework for fairness rules based on different fairness definitions, including Demographic Parity, Equal Opportunity, Equalized Odds, and Generic Fairness. The findings suggest that LLMs can be prompted to incorporate fairness considerations into their responses, but they still exhibit biases towards certain demographic groups. The study highlights the importance of fairness in AI systems and the need for further research to improve the fairness of LLMs.This paper explores the potential of Large Language Models (LLMs) to achieve fairness in classification tasks through in-context learning. The study investigates whether LLMs can incorporate fairness considerations into their responses when prompted with specific fairness rules. The research focuses on assessing the fairness of LLMs in classification tasks, particularly in the context of the UCI Adult Income Dataset, which contains demographic and financial information of individuals. The study compares the performance of three LLMs: GPT-4, LLaMA 2, and Gemini, in terms of accuracy and fairness metrics across zero-shot and few-shot learning setups. The results indicate that GPT-4 performs the best in terms of both accuracy and fairness, while LLaMA 2 and Gemini show suboptimal performance. The study also introduces a framework for fairness rules based on different fairness definitions, including Demographic Parity, Equal Opportunity, Equalized Odds, and Generic Fairness. The findings suggest that LLMs can be prompted to incorporate fairness considerations into their responses, but they still exhibit biases towards certain demographic groups. The study highlights the importance of fairness in AI systems and the need for further research to improve the fairness of LLMs.