She was beautiful.

She had thousands of followers, a carefully curated Instagram account, and the sort of flawless complexion that makes dermatologists question their career choices.

There was just one small detail.

She wasn't real.

Well, not real in the traditional sense.

No childhood photographs. No embarrassing middle-school haircut. No actual memories of that breathtaking sunset she supposedly enjoyed in Santorini.

Just pixels, algorithms, and a remarkably convincing imitation of a person.

And yet people followed her. Liked her photographs. Commented on her life.

They paid attention.

But according to new research, there may be something that even the most sophisticated digital appearance cannot automatically manufacture.

The feeling that someone is actually there.

The People Who Aren't People

Welcome to the increasingly strange world of virtual influencers.

They pose for photographs. Promote products. Build audiences. Develop recognizable personalities. Some are obviously computer-generated. Others could pass for the person sitting next to you at Starbucks.

Except the person at Starbucks probably can't spend the morning in Paris, the afternoon in Tokyo, and the evening watching a sunset in Bali without leaving their computer.

Virtual influencers can.

For brands, the appeal is obvious. Their schedules are limitless, their appearances carefully controlled, and their public personalities don't have to contend with the inconvenient realities of being alive.

But there's a difference between attracting an audience and making that audience feel something.

Researchers Kathleen Jagodnik and Alon Bartal at Bar-Ilan University wanted to investigate that difference.

Instead of asking people whether they liked artificial influencers, they examined how people actually responded to them.

And nearly 157,000 Instagram comments later, they found something worth talking about.

The Comments Told a Different Story

Their study, published October 29, 2026, in Scientific Reports, analyzed 156,918 comments directed toward human and AI virtual influencers.

Those comments had collectively received more than 12 million likes.

The researchers examined emotional expression, language patterns, and engagement.

Among comments directed toward human influencers, 41.5 percent expressed positive emotion.

For AI influencers, the figure was 24.8 percent.

Perhaps surprisingly, negative emotional expression was also lower: 5.8 percent for artificial influencers compared with 9.3 percent for humans.

People weren't simply being meaner to the machines.

They were expressing less emotion in either direction.

The researchers also identified differences in the richness and organization of the language people used.

Now, before we conclude that humanity has collectively decided to give its artificial celebrities the silent treatment, there are some important qualifications.

The study analyzed public comments, not private feelings. It couldn't establish that knowing an influencer was artificial caused the differences. And individual AI influencers varied substantially.

A comment section is not a laboratory instrument for measuring love.

Still, the pattern is intriguing.

We may be getting remarkably good at manufacturing attention without fully understanding what creates connection.

Perfection Has Its Problems

Think about the people you feel closest to.

Chances are, their appeal isn't limited to perfect photographs, carefully selected outfits, or the ability to deliver a flawless smile on command.

You've seen them tired.

You've heard them laugh at something ridiculous.

Maybe you've watched them struggle through a difficult season. Or you've been the person they called when everything fell apart.

Human relationships accumulate meaning through experiences that aren't always attractive, convenient, or predictable.

A virtual influencer can portray those experiences.

But portrayal and experience are different things.

Does an audience respond differently when it knows the person in the photograph has never actually lived through the story being told?

The Instagram study doesn't answer that question directly.

Fortunately, other researchers have been asking related questions.

And their findings complicate the picture.

Apparently, Artificial People Can Be Charming

A separate 2026 review examined 210 experiments involving virtual influencers.

The researchers found that artificial influencers generally generated more novelty than human influencers, while facing disadvantages in credibility and certain persuasive outcomes.

Interestingly, they were comparably effective at driving consumer engagement overall.

So the machines weren't necessarily losing the popularity contest.

They were playing a somewhat different game.

Then there's another 2026 study involving 540 young adults familiar with virtual human influencers.

Researchers examined how artificial personalities' self-disclosure—sharing seemingly personal information—related to emotional attachment.

The results were anything but straightforward.

Self-disclosure generally worked against the formation of parasocial relationships unless the virtual influencer was perceived as physically attractive. Perceived humanness also strengthened the relationship between those parasocial bonds and emotional attachment.

In other words, people can develop attachments to artificial personalities. But the conditions under which those attachments form are more complicated than simply making a digital face look realistic.

And then comes an especially interesting twist.

Researchers studying real social media influencers found that when those influencers delegated some replies to AI agents, followers who received the automated responses subsequently commented more often.

The responses were particularly effective when they were timely, relevant, and consistent with the influencer's style.

Of course, more comments don't prove deeper emotional relationships.

But they do suggest something important.

People respond to being responded to.

Which brings us to a distinction that may matter far more than whether an artificial face is beautiful.

What Happens When the Picture Talks Back?

A virtual influencer posting carefully curated photographs is one experience.

A conversational digital human that remembers what you said yesterday is another.

Imagine an older adult speaking with a digital companion that remembers the names of her grandchildren.

Or a teenager practicing a difficult conversation with a patient, responsive virtual character.

Or someone preparing for a medical appointment by asking questions of a digital concierge before speaking with a clinician.

None of those experiences is equivalent to following a fictional celebrity on Instagram.

They involve interaction.

Sometimes repeated interaction.

And potentially, a growing sense of familiarity.

That doesn't establish that the machine has feelings, understands the person as another human would, or experiences a relationship of its own.

But it raises an uncomfortable question.

If someone feels heard, does the value of that experience depend entirely on whether the listener is human?

Does Connection Require Someone on the Other Side?

Consider two conversations.

In the first, you're speaking with a person who barely listens, forgets what you've said, and spends most of the conversation waiting for their turn to talk.

In the second, you're speaking with a digital human that remembers your previous conversations, asks thoughtful questions, and helps you work through something difficult.

Which experience would leave you feeling more understood?

The answer may not be as obvious as we'd like.

But there's a distinction worth protecting.

Feeling understood doesn't necessarily mean someone understands you in the same way another person can.

And receiving a thoughtful response doesn't mean the system has lived through anything remotely like your experience.

We already form meaningful attachments to things that cannot reciprocate our feelings.

A favorite song. A childhood home. A fictional character whose story somehow helped us understand our own.

The feelings are real.

The relationship is different.

Artificial intelligence adds a new dimension because it can respond to us personally, often in ways that feel surprisingly attentive.

That's a powerful capability.

And power brings responsibility.

The Responsibility of Building Something That Feels Human

The more convincing digital humans become, the more carefully we need to think about what we're asking them to do.

A system designed to help someone practice a conversation is different from one designed to become their primary emotional relationship.

A companion that encourages an isolated person to reconnect with family and friends serves a different purpose from one that rewards the person for spending every available hour talking to the machine.

Both might generate impressive engagement numbers.

Only one is necessarily moving the person toward a healthier life.

That's why transparency matters.

People should know when they're interacting with artificial intelligence. They should understand what the system remembers, what it can and cannot do, and when human involvement is necessary.

Especially when the people using these systems are lonely, vulnerable, or seeking guidance about consequential decisions.

We should also resist the temptation to judge digital humans solely by how convincingly they imitate us.

Perhaps the better measure is whether the interaction actually helps the person on the other side of the screen.

The Missing Ingredient

The researchers examining nearly 157,000 Instagram comments didn't prove that artificial influencers are incapable of creating emotional responses.

They found that the emotional patterns surrounding them differ from those surrounding human influencers.

Other studies show that artificial personalities can attract engagement, inspire attachment, and influence behavior.

Taken together, the research leaves us with something more interesting than a simple verdict about whether AI can replace human connection.

It suggests we still have a great deal to learn about why people connect in the first place.

A realistic face isn't automatically a relationship.

A million followers aren't automatically a community.

And a machine that remembers your name isn't automatically your friend.

But a useful, thoughtful interaction with technology doesn't become worthless simply because no human being is sitting behind the screen.

Perhaps the most successful digital humans won't be the ones that convince us they're people.

Perhaps they'll be the ones that help us become a little more human ourselves.

After all, the real achievement of technology shouldn't be making machines look more like people.

It should be helping people live more meaningful lives.

And that leaves us with one final question.

If artificial intelligence can make us feel connected, will that connection bring us closer to one another—or simply closer to the machine?