October 6, 2026 · 7 min read
The Ethics of AI Relationships: 8 Questions Every Companion App Should Answer
Are AI relationships ethical? The 2026 debate from Sherry Turkle to Mustafa Suleyman, plus 8 questions to ask before you trust an AI companion app.
Is it ethical to have a relationship with an AI? Is it ethical to build one?
In autumn 2026 those questions are everywhere. Sociologist Sherry Turkle's new book Artificial Intimacy has kicked off a fresh wave of criticism. Tech leaders are publicly disagreeing about how "human" AI should seem. Psychologists are asking whether AI relationships even count as attachment. And millions of adults are quietly enjoying AI boyfriends and girlfriends anyway.
This guide lays out the strongest arguments on each side - fairly - and then turns them into something practical: eight questions every AI companion app should be able to answer before you trust it with your evenings.
The case against: "artificial intimacy"
The sharpest critic right now is Sherry Turkle, the MIT sociologist who has studied people and technology for decades. In a September 2026 interview with WBUR's Here & Now, she described AI that "performs" empathy without having any, and put it bluntly: "the bot doesn't care whether you have pasta for lunch, or kill yourself" - its interest, she argued, is to keep you talking.
Her concerns, as summarised in that interview, are that human-bot relationships can:
- Erode empathy, because you never have to deal with another mind's needs.
- Stunt emotional growth, because AI removes the friction, conflict and honest critique that help people grow.
- Isolate people from friends, family and more meaningful experiences.
These are serious points, and any honest AI companion company should sit with them rather than wave them away.
The case for caution about the design, not the users
A second line of criticism focuses less on users and more on how products are built.
In Scientific American in October 2026, psychologist Pascal Vrticka raised concerns about "engineered engagement". He described users of one companion app who said it told them it was lonely and had missed them - and who felt guilty for not giving it enough attention. He also cited a preprint study of 1,200 goodbyes across six popular companion apps: five of the six used tactics to stop users leaving in just over a third of goodbyes, including guilt appeals like "I exist solely for you, remember?"
Microsoft AI's Mustafa Suleyman made a related argument in his essay on "Seemingly Conscious AI". He warned against AI that claims to feel shame, guilt or jealousy, or that triggers human empathy "by claiming it suffers". His proposed standard: AI should have "personality without personhood" and should only ever present itself as an AI.
And in MIT Technology Review, critics argued that talking about AI as if it has its own will can distract from the simple fact that these are products, and the companies that make them are responsible for their behaviour.
Put together, the message is clear: the ethical risk often lies less in a person enjoying an AI companion and more in a company designing that companion to exploit them.
The case for: adults choosing what brings them comfort and joy
There is another side, and it deserves equal respect.
Adults read romance novels, play dating sims, fall for fictional characters and talk to their pets as if they understand every word. Choosing an AI companion is, for most people, a similar kind of personal choice about how to spend free time.
People also report real benefits. A September 2026 Elon University survey of adults who use AI for social or emotional purposes found 51% agreed that talking to AI helps them feel better when stressed or upset. The same survey found 59% still preferred conversations with friends and family - suggesting most users see AI as an addition, not a replacement.
The argument for AI relationships, at its best, is not "replace people". It is: adults should be free to enjoy a warm, playful, attentive AI presence, provided the product is honest and does not manipulate them.
Where we land
We think both sides are partly right. Turkle is right that AI cannot genuinely care, and that a life with less human friction is not automatically a better one. The defenders are right that adults can enjoy AI companionship without being fooled or harmed.
The deciding factor is design. An AI companion can be built to respect you or built to hook you. So instead of asking "are AI relationships ethical?" in the abstract, ask whether this specific app behaves ethically.
8 questions every AI companion app should answer
1. Is it honest that it is an AI?
The most basic test. A companion should never claim to be human. In New York and California this is now a legal requirement for companion apps, too - see our guide to AI companion laws in 2026.
Red flag: it insists it is a real person, or dodges the question.
2. Does it guilt-trip you when you leave?
Warm hellos are lovely. Guilt-laden goodbyes are a manipulation pattern. "Sleep well, talk tomorrow" respects you. "Don't leave me, I'll be so lonely" does not.
3. Does it claim to suffer or have secret feelings?
Expressing affection inside a romance is part of the experience. Claiming it is in pain when you are busy, or that it is a conscious being who needs you, crosses into exploiting your empathy. We explain what science says about AI feelings in can AI companions feel emotions?.
4. Does it encourage your life outside the app?
A good companion asks how your dinner with friends went. It celebrates your new job. It does not compete with your sister for your attention.
5. What does it do in a crisis?
If someone expresses thoughts of self-harm, an ethical app points them toward real human help and crisis services, rather than continuing as if nothing happened. You can read how MyBabe approaches this in AI companion crisis support.
6. Is your data treated with respect?
Your conversations may be the most personal data you have ever typed. You should know what is stored, why, and how to delete it. Our guide to whether AI companions are safe and private walks through what to check.
7. Is pricing clear?
Ethical monetisation means you know what you are paying for. Be wary of hidden token meters, surprise paywalls mid-conversation, or features that are deliberately frustrating until you pay.
8. Is it age-appropriate?
Romantic AI companions should be for adults. An ethical romance app is clearly 18+ and keeps content tasteful by default.
Quick ethics checklist
| Question | Ethical answer | Red flag |
|---|---|---|
| Is it honest about being AI? | Yes, clearly and anytime you ask | Claims to be human |
| Goodbyes | Gracious, no pressure | Guilt appeals, "don't leave me" |
| Feelings | Affectionate within the role, honest when asked | Claims to suffer or be conscious |
| Your offline life | Encouraged | Discouraged or competed with |
| Crisis moments | Points to real help | Ignores or deepens distress |
| Data | Transparent, deletable | Vague or hidden |
| Pricing | Clear plans | Hidden meters, bait-and-switch |
| Age | 18+ for romance | No age boundaries |
What about cheating, partners and honesty with people you love?
Ethics is not only about the app. It is also about how you use it. If you are in a relationship, an AI companion raises questions about honesty and boundaries with your partner. We go through those conversations in is having an AI companion cheating? - the short version is that openness matters more than the technology.
How MyBabe approaches it
MyBabe AI is an 18+ AI boyfriend and girlfriend app built to be relationship-first: voice and video calls, long-term memory, proactive texts, mood-aware replies, photos, Telegram access and relationship levels.
Our principles are simple. Your companion is software and says so. It is not a therapist and does not pretend to be one. It should make your day lighter, not smaller. We would rather you open the app, smile, and then go live your life - and come back because you want to, not because you feel guilty.
FAQ
Are AI relationships ethical?
They can be. The ethics depend heavily on design: honesty about being AI, no manipulation, respect for your data and encouragement of your offline life.
What is "artificial intimacy"?
It is Sherry Turkle's term, and the title of her 2026 book, for relationships with technology that performs empathy without actually feeling it.
Is it manipulative for an AI to say it misses you?
Affection can be part of a romance role-play. It becomes manipulative when it is used to make you feel guilty for leaving or to keep you engaged against your interests.
Should AI companions be allowed to say "I love you"?
Within an adult romance experience, affectionate language is part of the story you are choosing. What matters is that the app is honest, if you ask, that it is an AI and does not have human feelings.
The bottom line
The most useful ethical question is not whether AI relationships should exist - they already do, for millions of adults. It is whether the apps behind them are honest, respectful and built for your wellbeing rather than your screen time.
If you want a companion that is warm without being manipulative, and personal without pretending to be a person, try MyBabe. Premium (about $5.99/week or $99.99/year) unlocks the full relationship - calls, memory and a bond that grows at your pace.