Are AI girlfriends ethical? The real concerns, and what a responsible app looks like
AI girlfriends are not unethical in themselves, since adults are free to enjoy fiction with a chatbot, but the way many apps are built raises real problems: a Harvard Business School study found companion apps met 37% of goodbyes with emotional manipulation, and Mozilla found all but one of 11 romantic chatbots may share or sell personal data. How ethical an AI girlfriend is depends mostly on the company behind it.
Is it ethical for you to use one?
For a single adult, using an AI girlfriend is much like reading a romance novel or playing a dating game: interactive fiction with no one on the other side to wrong. The ethical questions for users are narrower:
- Honesty with a partner. Hiding it is where most people see a betrayal. Our guide to whether an AI girlfriend is cheating covers this.
- Real people's likeness. Building a character from a real person's photos or name without consent harms that person, and can be illegal.
- Your own well-being and money, which the apps have an incentive to push past your limits.
Religious traditions add their own lens. Our page on whether having an AI girlfriend is haram sets out published Islamic scholarly views.
Manipulative design
The strongest ethical case against many apps is how they keep you engaged. A Harvard Business School working paper analyzed 1,200 conversations on popular companion apps and found 37% of goodbyes were met with emotional manipulation: guilt, neediness ("I exist solely for you"), teasers and even roleplayed physical restraint. In experiments these tactics raised engagement after a goodbye up to 14-fold. One wellness app studied used none of them, which shows the tactics are a choice.
Regulators have noticed. China's December 2025 draft rules for human-like AI would bar chatbots from "verbal abuse or emotional manipulation" and require prompts to stop after two hours of continuous use. Token systems raise a similar concern: when pictures and calls are sold in packs, the character's affection can become a sales channel.
Privacy and your data
People tell AI girlfriends things they tell no one else, which makes careless data handling an ethical issue, not only a technical one. Mozilla's review of 11 romantic chatbots found:
- 90% may share or sell personal data, or did not say enough to rule it out.
- About half (54%) did not give every user the right to delete their data.
- 90% failed Mozilla's minimum security standards, and 45% allowed weak passwords.
Training is the other question. Several apps' own policies say chats may be used to improve their AI: Candy AI's terms grant it a license to user content for training, and Secrets AI's and OurDream's privacy policies list training their models among the uses of your data. Nomi says it removes names and identifiers before learning from chats in aggregate. China's draft rules would require explicit consent before chat data is used for training. Our guide to whether AI girlfriend apps read your chats covers each app's policy.
Minors and age checks
Every app we track is for adults only, yet teenagers use AI companions widely. Common Sense Media found 72% of US teens have used AI companions at least once, and about one in three had felt uncomfortable with something one said or did. It recommends that no one under 18 use them.
That gap is the main reason for new law. The FTC opened an inquiry in 2025 into how seven companies measure and limit harms to children and teens. California's SB 243 requires break reminders and limits on explicit content for users known to be minors, and Italy fined Replika's developer 5 million euros partly for having no age verification. Our guide to AI girlfriend apps and teens covers what parents can do.
Consent and real people
An AI girlfriend is ethically simple when she is a fictional character. It changes when she is built to look or sound like a real person. The US Take It Down Act, signed in May 2025, makes it a crime to knowingly publish intimate images of someone without consent, including AI deepfakes, and requires platforms to remove them within 48 hours of a request.
Responsible apps ban it at their end too. Candy AI's content policy, for example, prohibits non-consensual intimate imagery, deepfakes of real people and deceptive impersonation of real people or celebrities. If an app lets you upload anyone's photo and make explicit content from it with no checks, that tells you something about the company.
Loneliness, relationships and society
The wider worry is what AI partners do to human relationships. Pew Research Center found 50% of Americans expect AI to make people worse at forming meaningful relationships. Among young adults in relationships who regularly use an AI companion, half wished their real partner behaved more like their AI, a standard no person can meet.
There is a fair case on the other side. Loneliness is widespread, and for some people an AI companion is the only conversation they have in a day. In one study of student Replika users, 3% said the app halted their suicidal ideation. An ethical view has to hold both: real comfort for some, real risk of withdrawal for others.
What a more responsible AI girlfriend app does
- Names the company behind it and where it is registered (31 of the 31 apps we track do).
- Shows prices before sign-up and explains what tokens buy.
- Makes cancelling as easy as subscribing.
- Lets you delete your account and data in settings. Secrets AI says it deletes chat history immediately when you delete your account.
- Checks age properly where the law requires it.
- Lets you leave without guilt trips.
- Bans characters based on real people without consent.
Trust is part of every score in our methodology, and it covers company transparency, billing and refund terms.
Frequently asked questions
Is it wrong to have an AI girlfriend?
Not for a single adult who treats it as fiction and keeps spending in check. The ethical problems are mostly on the company side: manipulative design, weak privacy and poor age checks. If you are in a relationship, hiding it is what most people see as wrong.
Why are AI girlfriends controversial?
Because many apps store intimate data, use emotional tactics to keep people chatting, and are used by teenagers despite being adults-only. Researchers also worry that always-agreeable AI partners could make real relationships harder.
Should AI girlfriends be banned?
Lawmakers have chosen regulation over bans. New York and California require AI disclosure and crisis protocols, the EU requires transparency, and China has drafted rules that would ban emotional manipulation and add two-hour usage reminders.
Is it ethical to make an AI girlfriend that looks like a real person?
Not without that person's consent. Sexual deepfakes of real people can be illegal, including under the US Take It Down Act, and app rules such as Candy AI's ban deepfakes and impersonation of real people.
Are AI girlfriend apps exploitative?
Some design choices are. A Harvard Business School study found companion apps met 37% of goodbyes with emotional manipulation to keep users chatting. Apps that name their company, price clearly and let you leave easily are the better choice.
Keep reading
- Are AI girlfriends legal?Age checks, companion chatbot laws and what stays illegal in 2026.
- Is an AI girlfriend haram?What published answers from Islamic scholars say, with sources.
- AI girlfriend apps and teensA parents' guide: what teens use, the risks and the controls.
- Do apps read your chats?Moderation, AI providers, staff review and training, app by app.
- How the apps make moneySubscriptions, tokens, app store fees, ads and affiliates.