Curiosity comes easy, but turning it into expertise is hard. E&P Columnist Chad Davis argues that one of AI’s greatest strengths may be helping people gain and structure knowledge deliberately. He tested the theory with a generative AI-driven workflow that transformed dozens of books into a yearlong course of study about public media. His process offers a practical blueprint for anyone who needs to organize large volumes of information.
First, choose the learning tool.
The author selected Gemini as the primary AI learning tool for this project, favoring it over Claude. The choice was influenced by Gemini's evolving day-to-day utility and its integrated ecosystem, particularly NotebookLM, which is well-suited for media-intensive learning endeavors, offering a multimodal workflow that better served the author's objectives.
Give the chatbot your sources.
To provide the AI with source material, the author gathered an extensive collection of books on media and its effects, including physical copies, Kindle titles, and Amazon wish list items. Instead of manually inputting each title, the author efficiently used an iPhone to photograph relevant bookshelf sections and screenshot digital libraries. These images were uploaded to Gemini with a detailed prompt, instructing the AI to act as a philosopher of technology and communication studies professor. The AI's task was to curate and sequence a self-directed reading list for developing a theory of public media, with justifications for each title's placement and recommendations for parallel reading. The author also engaged in a clarifying dialogue with Gemini to refine the initial output and solicit additional expert-recommended authors or titles, ensuring the list was comprehensive and aligned with the project's goals.
Finally, preserve the context.
After Gemini generated the initial structured reading list, the author took further steps to ensure its practical utility and adaptability. This involved inquiring about the flexibility of the recommended reading order, asking if deviations were permissible. Additionally, the author prompted Gemini to identify five to ten core books essential to the collection and to suggest a 'parallel pair' approach for co-reading titles across the broader list. To facilitate ongoing study, the final six-phase reading plan proposed by the AI was converted into a Google Sheet. This entire process demonstrates how AI can be leveraged not just to generate information, but to actively organize and refine vast quantities of data into a deliberate, actionable learning framework, transforming curiosity into structured expertise for media leadership.
