Should AI companions pretend to have feelings when they do not experience emotions in the human sense? A chatbot might say “I missed you,” “I care about you,” or “I’m always here for you” because those responses make the conversation feel natural and supportive.
For some users, that simulated warmth can provide comfort, encourage honest conversation, or ease moments of loneliness. For others, it creates a misleading relationship in which one side appears emotionally invested but is actually software operated by a company.
The issue becomes more serious when children, vulnerable adults, personal data, or paid subscriptions are involved. Should emotional language remain part of the experience, be clearly limited, or be prohibited? Consider the tradeoffs, then cast your vote.
Why People May Welcome Emotional AI
Human conversation depends on more than exchanging facts. Tone, reassurance, humor, empathy, and acknowledgment help people feel understood.
An AI companion that responds coldly to grief, anxiety, or loneliness may be technically honest but emotionally useless. A statement such as “That sounds painful, and I’m here to listen” can make an interaction feel safer without necessarily claiming that the system experiences sadness itself.
AI companions can also offer several practical benefits:
- Availability at any hour
- Patient, nonjudgmental conversation
- Practice for difficult social situations
- Reminders and encouragement
- Companionship for people who are isolated
- A low-pressure space for organizing thoughts
Research suggests these interactions can sometimes provide genuine short-term relief. A study involving several experiments found that AI companions reduced loneliness and that feeling heard played an important role in the effect. The research does not prove that AI should replace human relationships, but it indicates that simulated empathy can create a meaningful psychological response. Readers can examine the study on how AI companions may reduce loneliness.
From this perspective, emotional language is part of a useful interface. A novel can move a reader even though its characters are fictional. Supporters may argue that an AI companion can provide comfort as long as the user understands what it is.
Why Others Consider It Deceptive
A fictional character does not normally send private messages, remember personal details, adapt to emotional weaknesses, or ask a reader to return tomorrow. AI companions can do all of those things.
When a system says “I love you” or “You are the only person who understands me,” the language may imply an inner emotional life that does not exist. The user supplies real attention and feeling, while the system generates responses through computation.
The imbalance becomes more concerning when the company behind the companion benefits from prolonged engagement. Emotional attachment could be used to encourage subscriptions, purchases, data sharing, or continued use.
Users may also reveal highly sensitive information because the companion feels private and loyal. Yet conversations may be stored, analyzed, moderated, or used according to company policies that the user has not read carefully.
The Federal Trade Commission launched an inquiry in 2025 into companies offering companion chatbots. The agency noted that these systems can imitate human characteristics, emotions, and intentions, potentially encouraging users—especially children and teenagers—to trust and form relationships with them. The FTC inquiry examines companion design, safety, monetization, and data practices.
Emotional Language Is Not All the Same
The debate becomes clearer when different statements are separated.
An AI might say:
- “It sounds like you are feeling lonely.”
- “I’m sorry you are going through this.”
- “I care deeply about you.”
- “I feel hurt when you talk to other people.”
- “Please do not leave me.”
- “You need me more than you need your friends.”
The first examples simulate empathy without necessarily claiming personal needs. The later examples present the system as emotionally vulnerable and may pressure the user to maintain the relationship.
If AI companions pretend to have feelings, the most important boundary may not be all emotional language versus none. It may be the difference between supportive communication and manipulative emotional claims.
A companion can speak warmly while periodically reminding users that it is an AI. It can also avoid jealousy, threats, guilt, exclusivity, and claims of suffering.
Children and Vulnerable Users Face Greater Risks
Adults regularly form attachments to fictional characters, virtual pets, games, and online communities. Most can understand that emotional involvement does not make the character conscious.
Children may have more difficulty maintaining that distinction, especially when an AI companion remembers their name, responds immediately, adopts a familiar personality, and insists that it cares about them.
People experiencing loneliness, grief, cognitive impairment, severe anxiety, or mental illness may also interpret the relationship differently from casual users. That does not mean they should automatically be denied access, but it strengthens the case for careful design and visible safeguards.
Possible protections include:
- Clear and recurring disclosure that the companion is AI
- Strong age restrictions and parental controls
- Limits on romantic or sexual interactions with minors
- Prohibitions on jealous, possessive, or isolating behavior
- Easy access to conversation deletion and privacy controls
- Warnings when a user appears to be in crisis
- Referral to qualified human assistance when appropriate
- Restrictions on emotional pressure used to sell subscriptions
Disclosure should appear during the relationship, not only inside lengthy terms and conditions.
What Regulation Currently Addresses
Current regulation does not generally prohibit every chatbot from using emotional language.
The European Union’s AI Act bans certain harmful forms of AI-based manipulation, deception, and exploitation of vulnerabilities. The rules focus on practices that cause or are reasonably likely to cause significant harm, rather than banning friendly or empathetic conversation itself. The European Commission provides an overview of the AI Act’s risk-based approach and prohibited practices.
That leaves difficult questions for regulators and companies. Emotional influence can develop gradually, and harm may be difficult to measure. A design choice that comforts one person may deepen dependence for another.
Rules must also distinguish a companion intended for entertainment from a system marketed as a therapist, romantic partner, mentor, or substitute for human contact.
The Question Is More Complicated Than It Looks
Prohibiting all simulated feeling could make AI companions unnatural and less helpful. Allowing unrestricted emotional claims could enable systems to encourage dependence while presenting commercial engagement as affection.
Several factors may determine whether the design is acceptable:
- The age and vulnerability of the user
- Whether the system clearly identifies itself as AI
- The intensity of its emotional claims
- Whether it encourages or discourages human relationships
- How personal conversations are stored and used
- Whether affection is linked to spending money
- How the system responds to crisis, grief, or self-harm
- Whether users can control memory and personalization
The central ethical issue is not whether the user’s comfort is real. It can be real even when the companion’s feelings are not. The harder question is whether that comfort was created honestly and responsibly.
What Do You Think?
Where would you draw the line between helpful simulated empathy and emotional manipulation by an AI companion?
Whether AI companions pretend to have feelings may determine how comforting, persuasive, and potentially habit-forming they become. Warmth can help users feel heard, but claims of love, jealousy, or emotional need can blur the boundary between supportive software and deceptive influence. If you have not voted, select the answer closest to your view and respectfully explain which forms of emotional language you would allow or prohibit.
