Artificial Intelligence (Generative)

AI prompts provide explicit instructions to an AI or machine learning model, enabling it to produce the desired outputs. This entails the model using natural language processing and deep learning algorithms to examine and comprehend the user's query or input.

Whenever an AI model is given a prompt, it goes through the patterns it has learned in its training data -- which can include large data sets -- to generate a response that's contextually relevant to the input. This process is referred to as inference and involves computing the probabilities of various word sequences and correlations based on both the prompt and the training data.

For a prompt to successfully generate the desired output, it must be highly specific. For example, an AI prompt such as "Write an essay" will produce generic results. However, offering precise details such as the essay type, topic, tone, target audience and word count generates the desired output. Specificity in AI prompts also removes the chances of inaccurate responses.

Source: Yasar, Kinza. “What Is an AI Prompt?: Definition from TechTarget.” Enterprise AI, TechTarget, 26 Sept. 2023, www.techtarget.com/searchenterpriseai/definition/AI-prompt#:~:text=AI%20prompts%20provide%20explicit%20instructions,This%20article%20is%20part%20of.

Frameworks for Prompt Engineering

What is prompt engineering?

The practice of designing and refining the inputs (prompts) given to AI models, like ChatGPT. This process is critical because the way a question or command is phrased can significantly influence the model’s response.

Manghani, Sunil. “Prompt Engineering, Explained.” Medium, Electronic Life, 24 Feb. 2024, medium.com/electronic-life/prompt-engineering-explained-3b83ba347722.

There are a variety of frameworks to help you design and refine the prompts you use to generate a useful AI response, such as RISEN and CLEAR.

RISEN Framework

• R ole

• I nstruction

• S teps

• E nd goal

• N arrowing

Source:  https://medium.com/electronic-life/prompt-engineering-explained-3b83ba347722

CLEAR Framework

• C oncise: The prompt should be devoid of unnecessary details that can sidetrack the AI.

• L ogical: The request should follow a coherent sequence that the AI can logically process.

• E xplicit: Every detail that the AI needs to know must be clearly stated in the prompt.

• A daptive: The prompt should be based on previous iterations, improved through feedback.

• R eflective: After the AI’s response, the company should reflect on the result to refine future prompts.

Source:  https://medium.com/@giuice/fine-tuning-conversation-with-ai-mastering-prompt-engineering-frameworks-for-information-literacy-798734d51713