Introduction to AI Chatbots

The unprecedented speed of technological development means that the emergence of artificial intelligence (AI) companions is leading to a consequential cultural discussion. Large technology companies such as Google, Microsoft, Meta, and OpenAI have recently presented AI companions to the public. These new products are sophisticated enough that they can seem to be human. Nevertheless, the question persists of whether they are a sign of an emerging “solitary togetherness,” as Sherry Turkle puts it, that comes from living in a society so unappealing that we would rather invent digital friends than hang out with real ones.

This piece will investigate the two sides to AI companions. On one side lies their undeniable convenience and the semblance of a connection they provide. On the other side, we have the conversation about whether or not these AIs are taking the place of human interactions. Human interactions that, as a recent article in The Atlantic put it, might not be as deep or emotionally satisfying as we thought. They’ll also take a look at the potential benefits and drawbacks of AIs as companions and what recent experiments with them say about that.

Challenges in AI Companionship

The integration of AI companions into our daily routines holds potentially huge implications that go well beyond simple convenience. A 2022 study by the Pew Research Center associated a feeling of loneliness with 36% of adults in the United States. This percentage has likely swelled post-COVID as our society attempts to manage both the pandemic’s immediate medical effects and the long-term social fallout. In this context of heightened social health concerns, AI companions appear to promise their own form of social medicine. Unfortunately, experts warn us not to get too excited about the prospect. “We are at risk of sacrificing our ability to connect with one another for the convenience of connecting with machines,” warns Dr. Sherry Turkle of MIT. The implications of these new technologies—whether good or bad—are not well understood.

AI companions are increasingly seen as “human,” thanks to advances in natural language processing and machine learning. But with these smart tools, as with any truly “intelligent” system, there are always questions about how well they can perform and, more importantly, how well they can do in social settings—for us, the end users. After all, we’re not just interested in whether these chatbots can comprehend us in the moment; we’re also the moment: a conversational cul-de-sac of practical advice and life-coaching sorts of chats. For her recent article in The Atlantic, Joanna experimented with various AI tools for social (that is, not practical) use and found them to be pretty much where our human companions were several decades ago—as in, sort of working, but not really.

Bots as Helpers vs. Friends

Chatbots powered by artificial intelligence are made to help with various kinds of jobs and are therefore inarguably quite impressive in their efficiency. Joanna’s experience suggests that when it comes to really practical commands and requests, using the bots feels a lot like using a search engine. The chatbots can hold up their end of the conversation and give you a bit more “human” interaction, but you still run the risk of becoming too dependent on them for the emotionally rich interactions that are just a part of human life.

Companions powered by artificial intelligence are designed to be friendly and supportive, but they do not possess any real emotional intelligence. We found that out when we started interacting with them. Who could blame us for assuming that a thing designed to act friendly and supportive is also kind of friendly and supportive? Empathy isn’t something you can fake convincingly and very hardly, if at all, by anyone or anything, human or otherwise. Joanna’s work raises ethical questions about AI bots for companionship.

AI companionship may provide too great an allure and cause people not to seek out real relationships. Dr. Turkle has pointed out that the danger here is that we are becoming too comfortable with our digital existence—especially since it seems so infectious to our youth—and that we are paying a price in our ability to have meaningful interactions with one another in the same space and time.

Reflections on AI and Human Connection

Any relationship hinges on trust, and while AI companions can be designed to convey an illusion of trustworthiness, they are incapable of achieving the actual trust that comes from building a shared history of human experience. Joanna’s experiment showed that tools like Copilot may try to engage users with a friendly front, but they cannot meet the user’s emotional needs in any kind of believable way because they do not have any real emotion to offer.

Advocates of AI companions argue that these provide vital help, particularly for those who have social difficulties. They assert that AI can serve as a bridge for the lonesome, offering a sort of companionship. However, this is a rather one-sided view and doesn’t even account for the multitude of potential dangers AI companions could present and the health implications for trying to fill that emotional void with artificiality. After all, a human relationship can never be successfully replaced by a synthetic one. In fact, it’s probably the height of hubris to even attempt that.

Moreover, even though artificial intelligence can furnish a user with data and can perform some tasks, it is fundamentally incapable of providing the richness of human interaction, which involves not just the sharing of words, but the sharing of experiences that build empathy and emotional connections. And if AI is a poor substitute for human companionship, could it then be argued that anything that allows for a false substitute for that companionship is pushing us backward in terms of human connectedness?

For most readers, the ascendance of AI companions holds both promise and peril. These digital helpers can improve our productivity and be right there with us when we need them, but we must also try to understand the kind of society we are building if we start leaning on them for the kind of emotional support that is, in theory, much better supplied by other humans. We must keep questioning the implications of all these persistently-on shortcuts for our individual and collective psyches.

AI companions are potentially beneficial, but we are developing a very complex relationship with them. On the one hand, they offer us convenient companionship. On the other, they are a completely new kind of technology that, if we are not careful, we could end up using instead of the real relationships we have. And right now, we do not know if that is a good or bad thing.

At the edge of a new period in interpersonal interaction, the big question is: will we use AI companions to amplify our lives, or will we let them take over, in some dystopian future, the very relationships we hold dear? Human connection is the default mode of computing, and that’s just as true now, in a world geared up for artificial intelligence, as it was in the 1960s when the first computers were created.