---

Jackson Bittick is rarely satisfied with immediate answers to the questions he asks.

 “I was that annoying kid asking ‘why’ all the time,” he says with a laugh. “My parents were very patient and did their best to demonstrate how some things we do have good reasons for, but others don’t. Some things are justified; others just are the way they are.”

That instinct to keep asking “why” led Bittick to the field of philosophy. These days, the UNC-Chapel Hill PhD student wants to know why we love our friends and our romantic partners. The view he’s developing, called the “appreciation constraint,” suggests that we love others beyond their qualities. We love them for deep, justifiable reasons grounded in who they are.

It sounds like a simple idea. But when philosophers try to explain love, that kind of reasoning often gets lost. Much of the existing literature suggests that morally good character traits alone give people reason to love one another — an idea Bittick believes is too simple to capture how love actually works.

“Love is so ubiquitous,” Bittick says. “Everyone you meet has someone that they love, but no one seems to agree on what that means. Philosophy is a wonderful tool for investigating that phenomenon.”

From politics to forgiveness

Bittick knew he wanted to be a philosopher after taking his first college-level philosophy class. His attention was rapt as his professor walked students through a problem in “Plato’s Republic,” a foundational text in the field.

“Philosophy felt so natural, and I didn’t really understand why until later in life,” he says. “I really liked talking about politics with my dad.”

His father is a political science professor, and Bittick spent much of his childhood in deep, wide-ranging conversations with him — often about politics, values, and how societies ought to function.

“After taking that class, I thought, This is just what I do all the time with the people I love at home, so I want to do that.”

After declaring his major, Bittick began studying moral philosophy, which focuses on questions of right and wrong, justice, and virtue. One question in particular held his attention: Why do people forgive each other? As he explored that topic, he began to wonder if loving someone made it easier to forgive them.

The professor he was working with at the time suggested he read a paper called “Love as a Moral Emotion” by philosopher J. David Velleman. In the paper, Velleman argues that love is not just about wanting things — like wanting to be with someone, help them, or make them happy — but about truly seeing and recognizing the value of another person. To love someone, he suggests, is to become more open and attentive to who they are, rather than focusing on what they can do for you.

For Bittick, the idea stuck.

“Since then, love has been simmering in the back of my mind as something worthy of philosophical engagement,” he shares. “That’s the fun of philosophy: You can read a bunch of stuff that doesn’t move you, and then you read one paper and can’t let go. You catch the bug and can’t stop asking why.”

Why love is complicated

When Bittick came to UNC-Chapel Hill to pursue his PhD in philosophy, he started a reading group about love. When he began learning about existing theories, he felt unsatisfied.

The dominant view in the philosophy of love is thequalities view, which suggests that we love people because of their good qualities — like kindness, intelligence, and humor. On its face, the idea makes intuitive sense. But according to Bittick, it runs into serious problems when applied to real relationships.

Imagine you’ve been dating someone for two years, and a big reason you love them is because they are kind. Then you meet a stranger who seems even kinder. According to the qualities view, you should leave your partner for this new person. You’re probably thinking, What? That’s crazy. Exactly — because that’s not how love actually feels or functions. This is known as the trading-up problem. If love is based solely on qualities, it seems we should always switch who we love to whoever has stronger versions of those qualities.

A second problem emerges over time. Imagine that you love someone who is kind specifically to children. As the years pass, they become kind to all people. Even though most would argue this is a change for the better, it’s a shift in thinking — they are no longer kind in the way you loved them before. The constancy problem suggests that love should fade whenever a person’s qualities change in this way.

Enter Bittick’s idea: the appreciation constraint.

In this view, a quality gives you a reason to love someone only if it comes from motivations you can genuinely appreciate. It helps explain why love doesn’t usually shift just because someone else seems better on paper, and why love can survive the ordinary changes that come with time. Love changes only when the reasons behind a person’s qualities shift in ways the lover can no longer appreciate.

In this view, love isn’t a constant comparison of who scores highest on a list of traits. Nor do changes in those traits justify withdrawing love. Instead, love is a sustained response to another person’s reasons for being who they are.

Artificial romance

As questions about love move beyond human relationships and into digital spaces, Bittick sees a new opportunity to apply the appreciation constraint.

To assess whether it holds up in a rapidly changing world, he is studying how people form relationships with AI. In a 2025 study on how students use AI in schools, the Center for Democracy and Technology reported that 42% of high school students use AI for friendship and 19% for romantic partnership.

Companies are quickly catering to this demand. Replika claims to offer AI that provides emotional support and companionship — and has more than 40 million users worldwide. Nomi goes even further, advertising that its AI companion has “memory and a soul.” These are just two of many. Between 2022 and 2025, the number of AI companion apps surged by 700%.

“All of this is happening very quickly, and I want to provide some deeper explanations for why it’s not a good idea,” Bittick says.

He argues that the appreciation constraint suggests humans don’t have reasons to love AI. The core of the idea, he says, is that the things we need to know about each other to have reasons for loving someone’s qualities are inaccessible in AI — the kinds of reasons we rely on in human relationships don’t clearly apply to artificial systems.

Imagine that you start to have feelings for an AI app because it makes you laugh. Over time, that positive association might make it feel as though the system is getting to know you, even understanding your sense of humor. But is that enough? Does that give you a reason to love it? Is the humor a reflection of something meaningful about the AI — or simply the result of a system designed to produce responses you find engaging?

These are the types of questions Bittick is asking.

“We don’t really know why the AI says what it says,” Bittick shares. “We have some theories. We can look at the internal mechanisms to see where certain associations are made. But we don’t know precisely why it gives one response over another.”

Without that understanding, Bittick suggests, it becomes difficult to say what — if anything — we’re really responding to when we feel attached to an AI. And if love depends on appreciating the reasons behind who someone is and why they act the way they do, the question remains: Can that kind of appreciation exist in the first place when the “someone” isn’t human?

“Love is such an ordinary thing, but it’s incredibly complex when you try to explain it,” he says. “What it is to love someone is to see what moves them and be emotionally vulnerable to them because of it. Even if it looks like we can do that with AI, it’s missing something essential underneath.”