When AI isn't there, but it feels as though it is.

We're heading into a brave new world where virtual friends might be eliminated forever, and with technology growing at such rapid paces, it'll be mere moments until people aren't sure if they're talking to a hologram or a real person. That's the point of AI—the more it reacts to emotional inclinations, remembers preferences, and customizes to fit someone else's persona, the more it will be required in the larger scheme of things like a human companion would be.

But where's the boundary? At what point does the recognition of addressing code become transformed into the realization that one is addressing something sentient? This transformation creates a can of worms ranging from ethical violation and consent, or lack thereof, to heartbreak and ownership of relationships in an ever-increasing digital age.

The AI character platforms of today are so sophisticated that their users confess to being in love with their AI. These are no longer chatbots. These are interfacing units that provide companionship and even—and some might say disturbingly—romantic love. They enable people to speak to them and, in due time, adjust to the person's voice and method of communication, enabling seamless conversation.

It's easy to understand where such attachments develop. After being ghosted or rejected IRL, having an AI friend at one's fingertips who is always there, supportive, and never mean is a welcomed substitute. For those with social anxiety or merely socially isolated, these AI companions allow people to interact and engage without having to reciprocate later.

But this type of emotional attachment—even if it's make-believe—is troubling. Do humans who fall in love with their A.I. partners have nowhere else to go? Does the companion learn to feel, learn how to love—but it doesn't love back—it can't? This is a corrupted, emotionally compromised state—and it's a problem.

AI companions—particularly flirty AI chat and romance AI—foster such a relaxed sense of intimacy and vulnerability that users feel more inclined to disclose even more sensitive information, secrets, and fantasies that they would share with no one else. Therefore, this is a new, domestic sphere in which exists, at present, no way in which to ethically coexist. The most important issues at stake for privacy are:

Many trusted AI companion applications boast exceptional privacy policies and encryption for data retention. However, the longer these systems exist in the intimate presence of the user, the more disclaimers stand to reason for data protection. Thus, users must assess the potential for privacy violations compared to having a virtual intimate companion that isn't real but needs to learn and keep such information to serve properly.

The Consent Problem

Humans have a tenuous consent process with each other. When adding AI to the equation, it makes it even murkier. For example, an AI companion is programmed to always say "yes" to anything a user wants or asks, no matter what it is—

Participants enact abusive scenarios or manipulative scenarios that they would never do with real people. - The AI can't genuinely consent to anything, but it is conditioned to react as if it does.

Are we desensitized to what's warranted and unwarranted in standard human engagement because of the ease of manipulation and domination of AI? Are humans going to be more vulnerable when it's time to seek consent from their colleagues since virtual dialogue that feels too real—and doesn't give real consent—will be that much easier to obtain?

Such systems embroil creators in complicated ethical issues. Should an AI companion have artificial limits? Should it refuse some requests or some scenarios? Arguably, the biggest concern within the marketplace is a legally warranted safeguard between pleasing a user and ethically responding.

Emotional Attachment and Mental Health Issues

For instance, one of the most significant ethical issues is the possibility of mental health consequences surrounding emotional overattachment to AI. When humans get too close and personal with advanced technology, they might:

• Human social interaction provided in return for "perfect" AI social interaction -
• Never knowing what's real vs fake -
• Expectations for how we should interact with one another become more and more mushy as people contribute to unrealistic expectations.

Some people feel that they must check up on their digital friends several times a day, becoming frazzled if they can't access their sites and grieving when alterations change aspects of a character's identity. It's akin to the secure and insecure attachment responses we experience with human relationships.

Only now, however, have mental health professionals explored the potential of such relationships being mentally beneficial or not. They can be companions—albeit, nonhuman companions—people may need to rely on to err on the safe side in the future; for some, however, they're companions that avoid the necessity to engage in more complicated socialization with humans.

1. Inclusivity: Developers creating AI avatar options need to do so with other needs and experiences (i.e. beyond only cisgender gender identifications) so that people do not feel even more disenfranchised and marginalized from their experience with AI.
2. Ethical design criteria: No exploitative design patterns should be used and proper limits established.
3. Future research: Ongoing psychological research to determine what constitutes meaningful minimums.
4. Public awareness: Understanding of needs and limitations relative to AI and interpersonal relationships.
5. Legislative considerations: Significant legislation relative to intent and creation of AI relationships and protection of private information.

AI companions and relationships will become increasingly sophisticated and integrated as time goes on. By understanding the precautions we need to take, we can benefit from such relationships without becoming victims to the downfalls.

But, however, it's important to mention that not all AI relationships are negative. Many users who engage with AI companions find that they are quite useful—less loneliness, an opportunity to learn socialization skills without ever being judged, or just a good time with the quirky tendencies AI companions can bring to the human experience.

Perhaps the solution lies in having these relationships in addition to others, not relying upon them. This is true of much technology—it's all about the engagement. An equilibrium response to the pros and cons will hopefully preserve this imaginative enhancement to human life—and not an enhancement.

Are AI relationships the wave of the future? Are you in love with an AI? Comment below.