Descriptions:
Google’s NotebookLM has become a popular research tool, but its detailed infographic output — generated by the Nano Banana model — frequently produces misspellings and completely jumbled text when multiple sources are selected. This video from Futurepedia documents a practical fix that sidesteps the platform’s lack of an in-app infographic editor.
The creator tests three approaches: Canva’s Magic Layers feature (works well on simple infographics but fails completely on complex multi-source ones, dropping text and changing fonts), iterative in-ChatGPT editing (fixes errors but tends to reintroduce artifacts into other sections when targeting specific corrections), and the recommended method — downloading the broken infographic and running it through ChatGPT with the prompt “Generate this exact infographic, but with every error in the text corrected.” This regeneration approach reliably resolves even heavily corrupted detailed infographics, as demonstrated side-by-side on a complex multi-source space exploration infographic where the original was nearly unreadable.
The video also notes a useful refinement: if a single regeneration pass still contains one or two errors, the fastest fix is to download the corrected version and run it through ChatGPT again rather than asking it to edit in place. The technique requires no paid tools beyond a standard ChatGPT account and takes under a minute per infographic.
📺 Source: Futurepedia · Published May 22, 2026
🏷️ Format: Tutorial Demo
