“Prompt engineering” refers to putting careful, deliberate thought into how you interrogate a GenAI tool.

Generic prompts tend to result in generic output featuring fluffy verbosity (lots of filler language). It also tends to be overly biased toward neutrality, providing factual (or at least factual sounding) answers but refrains from offering opinions or judgments. That part about "factual sounding": it can write plausible sounding answers that are in fact counter to fact. For an example of false facts, it can fabricate citations, such as using a real author's name and a real journal title but make up a title of an article the author plausibly could have written but never actually wrote.

But with thoughtful, well-designed prompts, and by following up initial responses by asking for further specific elaborations, you can vastly improve your results.

Here are some tips and strategies for effective prompt engineering:
 

• Define your objectives clearly – Decide what you want from the AI tool, and let it know. Do you want it to generate ideas? Answer a question? Summarize an article? Write a draft? 
• Provide specific context – Let the AI tool know the context of your question and the level of detail you want in your response. 
• Use clear and concise language – Unambiguous and straightforward instructions help to minimize misunderstandings and direct the AI tool more effectively.
• Use significant keywords – Include terms you would use searching for your topic in Google or a library catalog, to help the AI tool focus its response on the topic of interest.
• Set parameters and constraints – If you have specific requirements for the response (word count, style, format [e.g., paragraph, bullet points, chart]), let the AI tool know.
• Target audience – Tell the AI tool to respond to a fifth grader, a college graduate with no background in the subject, etc.
• Give examples – Useful if you want creative content or a specific format or style. For example, ask the AI tool to write a poem about bananas in the style of Lewis Carroll.
• Work in smaller chunks – Rather than ask it to "write a half hour speech on topic X," ask it to outline a half hour speech, then ask it to explain each of the individual points in its outline. You can then combine the separate points into a half hour speech.
• Iterative refinement – If the initial responses do not meet your expectations, refine your prompt based on the AI tool’s response, or explain what you would like adjusted or corrected.
• Experiment with different prompt styles – Try rephrasing your prompt, or using a different sort of prompt.
• Understand the AI tool’s limitations – An AI tool can only pull from what it has ingested, and only according to the algorithms etc. it runs on. They make mistakes, and are limited when dealing with highly specialized or nuanced queries.
• Verify any factual claims made by an AI tool before you use them – Since AI tools can make mistakes, check reliable sources to verify an AI tool's output.
• Privacy and confidentiality – Many AI tools ingest and remember queries and other inputs, so avoid including private, sensitive, or confidential information in your prompts.

GenAI tools can be helpful, but it takes work on your part to make sure they produce good work for you. You can make best use of a GenAI tool when you know enough about your topic to guide it to good results. A good general principle for academics using GenAI: outsource your work, not your thinking; i.e., use it as a tool to enhance your thinking.