ChatGPT: Issues of Omission and Silent Deletion during Document Updates

This repository investigates a critical issue observed in ChatGPT —
when updating documents via iterative dialogue, prior content may be silently omitted, truncated, or overwritten, sometimes without the user’s awareness.

---

📘 Repository Contents

🇬🇧 English

• Omission_Issues_in_Document_Update_via_ChatGPT.md

🇯🇵 日本語

• ChatGPTにおける文書更新時の削除・省略問題.md

---

🎯 Objective

To clarify the following:

• What was observed when documents were updated via dialogue
• Why it occurs, structurally and behaviourally
• How it affects document integrity, version control, and trust
• What technical or design-level solutions could prevent it

---

🌐 Intended Audience

• ChatGPT developers
• AI system designers concerned with output determinism and edit safety
• Users encountering loss of data or misalignment during document editing

---

🧪 Context

These insights are based on actual iterative interactions between I, as a user of ChatGPT and "Mike"(ChatGPT).
They highlight fundamental challenges in maintaining log integrity and user transparency in LLM-based content workflows.

---

✍️ Contribution Policy

Suggestions, discussions, or reproductions of the issue in other environments are welcome via GitHub Issues.