Skip to Main Content

Prompt Engineering for Academic Research

Introduction to Prompt Engineering for Generative Artificial Intelligence (GenAI) Academic Research at New York Tech.

ReAcT Prompting Framework (Read, Answer, Cite, and Think)

  • Why it's effective: ReAcT combines critical thinking, source validation, and logical reasoning, ensuring accuracy, depth, and transparency. It’s particularly powerful for academic research or tasks requiring factual precision.
  • Best Use Case: Suitable for researchers, academics, and any situation requiring high levels of critical evaluation and citation.
  • Example: "Read the latest studies on quantum computing algorithms. Answer what the key breakthroughs are. Cite your sources. Think about how these breakthroughs might impact cybersecurity."

CARE: Structure for Crafting AI Prompts (Context, Ask, Rules, Examples)

  • Why it's effective: CARE provides a balanced structure, ensuring prompts are clear, specific, and actionable while incorporating examples to guide the AI’s response.
  • Best Use Case: Ideal for tasks where clear guidance and contextual framing are crucial, such as user experience design, technical documentation, or precise task delegation.
  • Example: "Context: You are an AI assistant for a medical research lab. Ask: Provide a summary of the latest treatments for osteoarthritis. Rules: Keep the language simple, under 300 words. Examples: Past treatments included NSAIDs and physical therapy."

The "CLEAR" Framework (Clarity, Logic, Evidence, Action, Results)

  • Why it's effective: CLEAR emphasizes logical structure, actionable responses, and evidence-based reasoning, making it highly effective for technical fields where clarity and reliability are paramount.
  • Best Use Case: Particularly suited for engineering, biomedical research, or scenarios requiring methodical problem-solving.
  • Example: "Clarity: Provide a detailed explanation of the principle of photovoltaic cells. Logic: Ensure the explanation is step-by-step. Evidence: Reference key studies. Action: Suggest improvements in cell efficiency. Results: Describe potential future applications."

The C.R.E.A.T.E. Framework (Challenge, Role, Emotion, Ask, Tools, Examples)

  • Why it's effective: C.R.E.A.T.E. emphasizes creativity and engagement by framing the AI's "role" and infusing emotion into tasks, which can lead to more engaging and dynamic outputs.
  • Best Use Case: Useful for creative brainstorming, design-focused projects, and innovation-oriented research.
  • Example: "Challenge: Help architects design a net-zero energy building. Role: Assume you’re an environmental consultant. Emotion: Inspire architects to consider sustainability as a moral duty. Ask: Provide three innovative design concepts. Tools: Reference recent green-building case studies. Examples: Previous work includes solar-integrated facades."
© 2024 New York Institute of Technology