This journalist used AI to turn a 3rd grade journal into an engaging podcast

The emerging capabilities of AI tools are enabling creative new ways to revisit and analyze personal historical documents, as demonstrated by a writer’s experiment converting childhood diary entries into an AI-generated podcast.

The innovation process: Google’s NotebookLM and ChatGPT were combined to transform handwritten diary entries from third grade into an audio discussion format.

• ChatGPT Plus’s image processing capability was utilized to digitize photographs of the handwritten diary pages
• The transcribed text was then uploaded to NotebookLM, which generated a nearly 11-minute podcast featuring AI voices discussing the content
• NotebookLM, launched by Google in July 2023, added its Audio Overviews feature in September 2023, enabling podcast-style content generation from uploaded materials

Content analysis: The AI-generated podcast examined various themes from the diary entries, providing commentary on the writer’s childhood experiences and development.

• The discussion covered topics including the author’s pet caterpillar named Alfred, demonstrating early signs of nurturing behavior
• The AI identified and analyzed patterns showing curiosity and eagerness to learn
• Friendship emerged as a central theme in the diary entries, which the AI voices explored in detail

Technical limitations: While the experiment demonstrated impressive capabilities, several constraints and drawbacks became apparent during the process.

• ChatGPT’s free version has upload limitations, necessitating a $20 monthly Plus subscription for the image processing task
• The AI analysis tended toward superficial observations and showed repetitive patterns in its commentary
• The technology performed better as a document analysis tool than as a podcast generator

Educational potential: Despite its current limitations, the technology shows promise for educational applications.

• Students could potentially use the tool to create study materials from course content
• The question-answering capabilities of NotebookLM proved particularly useful for document analysis
• The technology demonstrates potential for making historical documents more accessible and engaging

Future implications: While this experiment reveals both the capabilities and limitations of current AI technology, it points toward increasingly sophisticated ways of interacting with personal and historical documents, though careful consideration must be given to the balance between novelty and meaningful analysis.