AI for Teaching and Learning |
Generating Prompts
John Behrens, director of the Technology and Digital Studies Program in the College of Arts and Letters, explains that when using text-to-text AI systems, how you prompt the system significantly impacts the results. Effective prompts should be specific and detailed, considering the system’s probabilistic nature. The CRAFT framework (Context, Role, Action, Format, Target) can help structure prompts effectively. Iterative prompting, providing examples, and asking the AI to evaluate or refine its output can enhance the interaction, as well. You can also experiment with different personas and languages to tailor responses. For complex tasks, guiding the AI to think step-by-step or take its time can improve the output, making AI a more effective tool in various contexts.