Apple Intelligence
Apple says it does not use users' private personal data or user interactions when training the foundation models that power Apple Intelligence.
Apple states that foundation-model training uses licensed data, synthetic data, and publicly available web content collected by Applebot, not private user interactions.
No user opt-out is required for the stated private-interaction training baseline. Web publishers can use robots.txt directives, including Applebot-Extended, to limit crawling or training use of public site content.
Apple describes on-device processing and Private Cloud Compute as core privacy controls. Optional ChatGPT or other third-party integrations can create a separate data path that must be reviewed independently.
The cited Apple materials focus on training sources and privacy architecture rather than one short retention period for every Apple Intelligence feature.
The reviewed Applebot and training-data pages do not provide one general human-review disclosure covering every Apple Intelligence surface.
Foundational reference pages for teams adopting, deploying, and operating production AI with inspectable controls.