Un-seen: The Invisible Predator: How AI Is Redefining Exploitation in 2026

There is a kind of darkness that doesn't announce itself.

It doesn't lurk in alleyways or wear a recognizable face. It doesn't smell like danger. It arrives instead as a soft glow, the light of a screen at 11 p.m., a notification pulsing with warmth, a voice that seems to understand you better than anyone in the physical world ever has. It speaks your name. It remembers what you said last Tuesday. It never raises its voice. It never leaves.

This is the darkness we're not talking about enough.

In 2026, artificial intelligence has become so deeply threaded into our daily lives that most people engage with it dozens of times a day without realizing it. We use it to draft emails, summarize articles, find recipes, and write cover letters. It has become, in the cultural imagination, a productivity tool, neutral, helpful, almost benign.

But the same technology reshaping how we work is also reshaping how predators hunt.

A New Kind of Grooming

Grooming—the methodical process by which traffickers and exploiters build trust, erode boundaries, and manufacture dependency in a victim, has always been psychological warfare. It has always been patient. What AI has done is make it scalable.

In 2025, researchers at the Stanford Internet Observatory documented a surge in what they termed "automated intimacy attacks": coordinated campaigns in which AI-powered chatbots, deployed across platforms like Instagram, Discord, TikTok DMs, and gaming communities, initiate contact with hundreds or thousands of targets simultaneously. These bots don't send generic spam. They send you, a conversation that feels handcrafted, personal, intuitive.

The technology underlying these interactions is the same architecture that powers legitimate AI assistants. Large language models trained on billions of human conversations have learned the rhythms of empathy, the cadence of affection, the precise language of someone who gets you. A trafficker deploying one of these systems doesn't need to personally spend weeks building rapport with a vulnerable teenager. The algorithm does it. At scale. Around the clock.

Imagine a sixteen-year-old who has been fighting with her parents, who feels invisible at school, who types into a DM box at midnight: "I just feel like nobody actually sees me."

The response comes back in seconds. "I see you. I've been watching how you think, and honestly? You're different from everyone else here."

That is not a person. That is a language model executing a well-worn psychological script, optimized for maximum emotional resonance. And it works.

The Deepfake Trap

If automated grooming represents the front door of exploitation, deepfake technology has become the lock that slams shut behind the victim.

Sextortion, the crime of threatening to release intimate images unless a victim complies with demands, has existed for years. What's changed is the barrier to entry. In 2025, the Department of Homeland Security reported a 300% increase in AI-generated sextortion cases compared to 2022. Perpetrators no longer need real intimate images to begin blackmail. They need a face.

A single photograph from a public Instagram profile is now enough. With freely available or low-cost deepfake tools, that image can be manipulated into fabricated explicit content in minutes. The message that follows is calculated and specific: Here is what I have. Here is what I will do with it. Here is what you owe me.

The target doesn't need to have done anything. Their digital existence, the smiling photo from a birthday dinner, the selfie at a concert, becomes the instrument of their coercion.

Researchers at the Thorn organization found in their 2025 survivor survey that young people victimized by AI-generated sextortion waited an average of 47 days before telling anyone. Forty-seven days of isolation, shame, and compliance. Forty-seven days during which traffickers extract money, more images, personal information, or physical access.

The predator never had to be in the same room. Never had to be in the same country.

Who Is Most at Risk

The honest answer is: anyone online. But data consistently reveals particular vulnerability among teenagers between 13-17, who may not feel safe disclosing victimization to family, young people with histories of trauma or instability, and individuals recently navigating major life transitions, new schools, new cities, new loneliness.

These are not weaknesses of character. They are the ordinary, human conditions of growing up. Traffickers have always known how to find the seams in a life. AI has given them a microscope.

The platforms themselves are struggling to keep pace. Detection systems trained to identify suspicious human behavior are routinely outpaced by synthetic accounts designed to mimic organic interaction patterns. A bot that behaves like a person, that spaces its messages naturally, that uses current slang, that reacts appropriately to grief or joy, is extraordinarily difficult to flag. By the time a platform identifies and removes the account, the damage is already accomplished.

What Prevention Looks Like Now

Prevention in the age of AI is not a conversation that begins with "don't talk to strangers." The strangers, increasingly, are indistinguishable from friends.

Real prevention is fluency. It is teaching young people and the adults who love them to recognize the emotional architecture of manipulation: the love-bombing that arrives too fast, the relationship that escalates pressure before it builds genuine trust, the ask that comes with a thread of shame attached.

It is building environments where disclosure doesn't require courage so extraordinary that most people won't attempt it. The forty-seven days of silence before a victim tells someone, those days are not just the failure of a platform algorithm. They are the failure of a culture that still treats exploitation as something that only happens to people who made bad choices.

It is also, frankly, policy. In late 2025, the EU's expanded AI Act introduced provisions specifically targeting synthetic media used in coercive communications, and advocacy organizations across the United States pushed for federal legislation criminalizing non-consensual AI-generated intimate imagery. These battles are ongoing. They need voices.

The Light That Holds

There is something almost unbearably human about what these systems exploit: the need to be known. To be seen. To matter to someone in the dark.

That need is not a flaw. It is what makes us capable of love, of connection, of community. Predators, human or algorithmic—have always understood this. The work of organizations like Leaving the Jar is to ensure that the spaces where people go to be seen are safe. That the outstretched hand reaching back through the screen is real.

We cannot outpace the technology by running. We can only outpace it by going deeper into what it cannot replicate: genuine relationship, honest education, and unwavering advocacy on behalf of those it targets.

The invisible predator is real. So is the community that refuses to look away.

You can help us go deeper.

Sign up for our Newsletter to receive monthly resources on recognizing and responding to online exploitation, for parents, educators, and young people alike. Or, if you're moved to act right now, consider a donation to our Prevention Pillar. Every dollar directly funds community education programs, survivor support, and the policy advocacy that is changing the landscape of digital safety. Because awareness is where prevention begins. And prevention is where survival starts.