Editorial correction — September 20, 2026. Withdrew inadequately established national romance and user-count estimates and claims that chatbot use necessarily displaces human relationships. Grounded the accountability questions in the FTC inquiry.
A conversational product can feel personal while remaining a commercial service. That difference matters most when a user begins sharing what they would normally tell a trusted friend.
The documented concern
In September 2025, the FTC announced a study of seven companies’ AI chatbots, examining subjects including risks to children and teens, disclosures, engagement incentives, and handling of conversational data. A study is not a finding that every company violated the law. Read the announcement and orders.
The previous article presented poorly established survey and market estimates as a national measurement of romantic attachment. It also inferred dissatisfaction in human relationships from chatbot use. Those inferences are not justified by the material established here and have been withdrawn.
The reciprocity problem
A human relationship involves another person’s needs and boundaries. A product’s availability can be useful, but availability alone is not reciprocity. Our concern is what happens when the language of care makes the business arrangement difficult to see.
This does not mean every user is isolated or incapable of distinguishing software from a person. It means a company should explain the arrangement clearly enough that someone seeking comfort can make an informed decision.
Five questions before trusting the interface
- Who can access the conversations, and under what conditions?
- What happens to the conversation history when an account is deleted?
- Which features are designed to extend engagement or sell an upgrade?
- How does the product explain its limits when a user seeks consequential advice?
- What protections apply to minors, and how are they enforced?
A reassuring persona does not answer those questions. The relevant evidence is the product’s policies, controls, and tested behavior. An ethical assessment should examine those features rather than use a dramatic user-count estimate as proof that human intimacy is disappearing.
The moral demand is transparency at the moment it matters: when the service feels most like a confidant and the user is least inclined to inspect the contract.