FOOD FOR THOUGHT

"Food for Thought is a bi-monthly newsletter from the Seaford High School Library that provides you with up-to-date information on best practices for AI usage, trending issues in research, and features a recommended read."

MACHINE LEARNING (ML) - learn from data and improve performance. ML identifies patterns, makes predictions, and refines decision-making based on real-world interactions.

NATURAL LANGUAGE PROCESSING (NLP) - understand, interpret, and generate human language. NLP allows the AI to process text and speech, making interactions more natural and intelligent.

GENERATIVE AI - creates new content—from text and images to videos and music to form unique, human-like outputs.

To make an AI say "I don't know" instead of generating false or misleading information, you can implement the following techniques:

1. Confidence Thresholding - prompt the AI model to generate responses with high accuracy threshold (e.g., 80%)—if the model’s confidence in an answer is below that, it responds with "I don’t know."

2. Source Verification - Require the model to retrieve answers only from trusted source. If no reliable source is found, the AI responds with "I don’t know."

3. Trust Layers & Citation-Based Responses - Implement a system where AI only answers if it can show a valid source. If no source is available, it defaults to "I don’t know."

Lately, I've been following Peter Diamandis. Peter has a background in molecular genetics, aerospace engineering, has founded over 25 companies in health-tech, space, venture capital, and is passionate about AI’s transformative role in education.

Here are some reading suggestions for anyone looking to dive into a good book this spring or summer.