Skip to content
Pie · AI · Studio
GUIDE · A-01 · ENArticle OnePublic draft1 min read

Do not touch my data

Data includes memory, work, project context, preferences, relationships, and model choice.

Do not touch my data is not just about files. It is about the memory, workflows, and context a person builds with AI. Those things should not become a platform’s permanent mine by default.

The danger is not only whether a platform is evil today. The danger is lock-in: the deeper you use it, the harder it becomes to move your memory, style, and projects elsewhere.

So this article asks for exportable data, portable memory, and replaceable models. Even a slower alternative keeps an exit open if it belongs to you.

Living water · external evidence

OPEN SOURCES

The article is the author's own judgement (bedrock). Below are independently verifiable external sources (living water): each notes its type, stability and the date it was last checked, so you can look and push back yourself.

  • ReferenceBedrock · stable

    POSSE — Publish on your Own Site, Syndicate Elsewhere

    IndieWeb

    POSSE is IndieWeb's model: keep the canonical copy on your own site, syndicate to platforms as copies. A concrete exit from handing your content and data to a platform.

    Last checked 2026-05-30

  • LawBedrock · stable

    General Data Protection Regulation (EU) 2016/679

    European Union

    GDPR turned personal control over one's own data into enforceable law — the most teeth-bearing real-world reference behind ‘don't touch my data’.

    Last checked 2026-05-30

  • LawLiving water · still moving

    Artificial Intelligence Act — Regulation (EU) 2024/1689

    European Union (in force 2024-08-01)

    The first comprehensive AI law, in force since Aug 2024 but phasing in past 2026. The rules are still landing — living water that needs re-checking.

    Last checked 2026-05-30

  • LawLiving water · still moving

    The New York Times Company v. Microsoft Corporation & OpenAI

    U.S. District Court, S.D.N.Y. (filed 2023-12-27)

    The Times sued OpenAI and Microsoft in late 2023 over unauthorized use of its articles for training — a landmark fight over who owns training data. Still ongoing; the outcome will move.

    Last checked 2026-05-30

Got a source to suggest?

Take it with you
Share on XWeibo

No tracking. For platforms without a web share button (Xiaohongshu, WeChat), copy the link or use your device’s share sheet.

Back to guide index