If you have not reread an AI companion app’s privacy policy since you signed up, there is a decent chance it has changed since then. Consumer AI products in this category revise their legal pages more often than most software categories, frequently without a public changelog or notification email.
What we know:
Regional AI-transparency rules — including the EU’s AI Act disclosure requirements EmberGF has tracked on this desk — are pushing companion operators toward clearer labelling of AI-generated content and data-use disclosures. Several companion platforms have added or expanded language around model training use of conversation data in their policies over the past year, a pattern we have documented directly in individual EmberGF reviews (see our Candy.ai and CrushOn.ai coverage).
What remains open:
EmberGF does not track every company’s policy diff history, so we cannot say how many operators have made a material change in any given month, or whether a specific company’s current policy differs from the version you agreed to. That is a per-account question only you can answer by comparing the current live policy against what you remember agreeing to.
Bottom line
Bookmark the privacy policy URL of any AI companion you use regularly and reread it every few months — not because something is necessarily wrong, but because “I agreed to this once” is not the same as “this is still what I agreed to.”