This paper explores the use of Artificial Intelligence (AI) to enhance the diversity and quality of idea generation in the context of creativity and innovation. While AI systems like GPT-4 can produce high-quality ideas, they often fail to generate a diverse set of ideas, which limits the novelty and overall quality of the best idea. The study investigates methods to increase the dispersion of AI-generated ideas, focusing on different prompting strategies and their impact on Cosine Similarity, the number of unique ideas, and the speed of idea exhaustion. The research uses GPT-4 to explore various prompting methods, including minimal prompting, persona-based prompting, creativity techniques, and Chain-of-Thought (CoT) prompting. The domain of idea generation is set for developing a new product for college students priced under $50. The findings show that: 1. AI-generated ideas, without special prompting, are less diverse compared to ideas generated by human groups. 2. Prompt engineering can significantly improve the diversity of AI-generated ideas, with CoT prompting leading to the highest diversity and the closest match to human-generated ideas. 3. CoT prompting not only increases diversity but also generates the highest number of unique ideas, from around 3700 to 4700. The study concludes that while AI can automate and enhance the idea generation process, prompt engineering, particularly CoT prompting, is crucial for generating more diverse and innovative ideas. The results have implications for anyone using AI to augment human idea generation, suggesting that longer and more elaborate prompts, especially CoT prompting, can significantly improve the diversity and quality of ideas.This paper explores the use of Artificial Intelligence (AI) to enhance the diversity and quality of idea generation in the context of creativity and innovation. While AI systems like GPT-4 can produce high-quality ideas, they often fail to generate a diverse set of ideas, which limits the novelty and overall quality of the best idea. The study investigates methods to increase the dispersion of AI-generated ideas, focusing on different prompting strategies and their impact on Cosine Similarity, the number of unique ideas, and the speed of idea exhaustion. The research uses GPT-4 to explore various prompting methods, including minimal prompting, persona-based prompting, creativity techniques, and Chain-of-Thought (CoT) prompting. The domain of idea generation is set for developing a new product for college students priced under $50. The findings show that: 1. AI-generated ideas, without special prompting, are less diverse compared to ideas generated by human groups. 2. Prompt engineering can significantly improve the diversity of AI-generated ideas, with CoT prompting leading to the highest diversity and the closest match to human-generated ideas. 3. CoT prompting not only increases diversity but also generates the highest number of unique ideas, from around 3700 to 4700. The study concludes that while AI can automate and enhance the idea generation process, prompt engineering, particularly CoT prompting, is crucial for generating more diverse and innovative ideas. The results have implications for anyone using AI to augment human idea generation, suggesting that longer and more elaborate prompts, especially CoT prompting, can significantly improve the diversity and quality of ideas.