Could weird recommendations mean an app is trying to mess with relationships?

Those examples can feel personal, but they don’t establish intent. Recommendation systems can make bad guesses, use signals that aren’t obvious to the person seeing the feed, or show stale and inaccurate activity records. A strange suggestion, or even a disputed record, isn’t enough to show that a company is trying to damage relationships or steer people toward AI companions.

I’d separate two questions: why these particular posts appeared, and whether there’s evidence of a bigger strategy. The first might be worth investigating through account activity, ad settings, connected devices, and login security. The second needs evidence linking the recommendations to a plan to promote AI companions. The feed alone can’t show that.

For the relationship side, I’d talk through the odd recommendations together and check the account settings, rather than treating them as evidence of someone’s behavior. Feeds are messy and sometimes unsettling, but a recommendation isn’t proof that a person searched for or signed up for something. The AI-companion theory is possible to imagine, but this account of recommendations doesn’t support it on its own.