A manuscript is more than a file — it’s intellectual property, often the result of years of work. Anyone entrusting it to an AI platform should know exactly what happens to that data. The uncomfortable truth: with many services this is alarmingly unclear.
The crucial question: training
The most important question is: are my texts used for training? Some services use submitted content to improve their models — your unpublished book could feed into others’ output. A serious provider rules this out: no use of your manuscripts for public training, contractually assured.
No training of public models on your data
Deletion on request, provable
Data export possible anytime
Transparent list of sub-processors
If a service doesn’t clearly say it won’t use your texts for training, you have to assume it does.
Where does your data live?
Data residency sounds technical but is legally decisive. If your manuscripts sit on EU servers, European privacy rules apply. If they vanish onto servers in unclear jurisdictions, you lose rights you may never get back. AES-256 encryption and EU residency aren’t extras but baseline.
Control is a feature
In the end it’s about control: over your content, your sources, your cover. A good service gives you that control fully — including the ability to delete everything again. Data privacy isn’t fine print but a mark of quality. Those who protect their manuscript protect their most important investment.