Article

Privacy

Ruben Hassid, How to AI · 2026-09-13.September 13, 2026

Summary

This article follows the “life of a prompt” to show what can happen after information leaves your laptop and reaches services such as ChatGPT or Claude: it may be stored, scanned for safety, sampled for review, retained in backups, or used for model improvement depending on the product and settings. The useful distinction is that a model does not simply “remember” your conversation, but the service around it may still retain and process the data. For knowledge workers, the practical lesson is to treat AI privacy as a matter of account type, settings, retention, sharing, and data sensitivity—not just whether another user can see your chat.

Article source handoff

Ruben Hassid, How to AI · 2026-09-13

This page is a short editorial source record. The summary and notes help you decide whether to continue, while the full article remains with the original author or publisher.

PublicationHow to AI
Author / creatorRuben Hassid
Source dateSeptember 13, 2026
Read full article at source

Key Takeaways

  • “The model forgets” is not the same as “the company deletes the data.” The article distinguishes the model’s temporary use of a prompt from the surrounding systems that may store conversation history, memory, safety records, backups, or training copies.
  • Privacy depends heavily on how the account is configured. The article argues that training settings, temporary/incognito modes, memory, feedback buttons, and whether you use a personal or business account can materially change how your conversations are handled.
  • Some of the largest privacy risks sit outside the core model itself. Shared links can expose conversations, browser extensions may gain access to chats, organisational accounts may make data available to workspace owners, and legal requests can create additional routes to disclosure.
  • Sensitive work should be governed by the data, not by convenience. The article’s practical recommendation is to use stronger business or API protections for client work, legal material, health information, or unreleased intellectual property rather than placing highly sensitive material into a personal AI account.