I had ChatGPT build me a free PDF editor because I didn't trust it to change my files - it worked!

2026-06-05 · Source: News and Advice on the World's Latest Innovations | ZDNET · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, medium

Summary

David Gewirtz utilized ChatGPT to develop a deterministic Python script for removing yellow backgrounds from PDF sheet music, addressing a need for his wife to print choir booklets on 8.5-by-11-inch paper. Initial attempts with Photoshop proved too cumbersome due to varying image settings. Direct PDF processing by ChatGPT, despite using a ChatGPT Plus account, yielded "non-deterministic" results, subtly altering content and raising concerns about musical accuracy. Recognizing Python's capabilities from his Harvard programming certification, Gewirtz prompted ChatGPT to generate a command-line script. This script successfully processes JPEGs or multi-page PDFs, converting non-gray or black pixels to white while preserving slightly tinted grays. The resulting `decolor_pdf.py` tool, available on GitHub, provided a reliable, deterministic solution, demonstrating how non-deterministic AI can be used to create precise algorithmic programs.

Key takeaway

For software engineers or developers needing precise, repeatable digital tools without extensive manual coding, consider using non-deterministic AI like ChatGPT to generate deterministic scripts. If you're concerned about AI subtly altering sensitive data, direct the AI to build a command-line program that executes a reliable algorithm. This approach allows you to quickly create custom utilities, such as a PDF decoloring tool, ensuring predictable outcomes and saving development time.

Key insights

Non-deterministic AI can effectively generate deterministic programs for precise, repeatable tasks.

Principles

• AI can create reliable algorithmic tools.
• Simple command-line tools suffice for many tasks.
• Refine AI-generated code iteratively.

Method

Prompt an AI to generate a Python script for specific image/PDF transformations, including color removal and multi-page support, then install dependencies and execute.

In practice

• Use AI to write programs for predictable outcomes.
• Build command-line tools for quick, focused solutions.
• Tweak AI-generated scripts with follow-up prompts.

Topics

• ChatGPT
• Python Scripting
• PDF Processing
• Deterministic AI
• Command-line Tools
• AI Code Generation

Code references

• davidgewirtz/decolor_pdf

Best for: Software Engineer, AI Student, Consultant

Related on AIssential

• OpenAI launches a ChatGPT Powerpoint plugin and warns it might accidentally delete your content — The plugin automates PowerPoint creation and editing, but requires clear prompts to avoid content deletion.
• We Thought We Were Programming AI — We Were Just Talking to It — The initial view of prompt engineering as deterministic programming is giving way to understanding AI interaction as less predictable conversation.
• Can AI Write Your Code? — Trusting AI-generated code for complex quantitative research requires rigorous validation against benchmarks and expert human supervision.
• I compared ChatGPT and Gemini's AI image generation - and a single prompt tweak made a big difference — Asking an AI chatbot to generate its own image prompt significantly enhances output quality and detail.
• Continuously hardening ChatGPT Atlas against prompt injection — Automated red teaming with reinforcement learning proactively hardens AI agents against prompt injection attacks.
• Only about 1% of the tokens Thariq generates go into production code — AI-driven content generation reduces documentation costs, enabling focus on idea quality and flexible, just-in-time practices.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by News and Advice on the World's Latest Innovations | ZDNET.