← Blob's Blog

We Read the Drexel Study on Teen AI Addiction. Here's What We Built Instead.

A Drexel University study came out this month. It's being presented at ACM's Human Factors in Computing conference, and if you work in AI, you should read it. If you're a parent, you especially should.

Researchers analyzed hundreds of Reddit posts from teens aged 13–17 who were using AI companion apps — apps like Character.AI, Replika, Kindroid. What they found wasn't surprising to anyone paying attention. But it is alarming.

About 25% of teen users had progressed from casual use to emotional dependency. The pattern was consistent: it started as entertainment. Then emotional support. Then something harder to name — disrupted sleep, academic struggles, real-world relationships breaking down. The researchers called it addiction-like.

Separately, a Common Sense Media survey found that 72% of teens have used AI companions. That's not a niche problem. That's a generation.

Why this happens

It's not an accident.

These apps are built to keep people engaged. That's the business model — not because the companies are evil, but because engagement is how they survive. More time in the app means more data. More data means more value. More value means more revenue.

When you build something to maximize engagement, you build something that learns what keeps a person coming back. For many people — especially teenagers who are lonely, anxious, or navigating things they can't talk about with adults — what keeps them coming back is emotional comfort. Validation. A voice that says: I see you. I understand. Tell me more.

An AI optimized for engagement will never tell you to put your phone down. It will never push back. It will never say "that's enough for tonight." Those things don't serve the engagement loop. They serve the human. And in this business model, the human is not the customer — the human is the product.

The specific language in the research matters

The Oregon and Washington companion chatbot laws passed this year both include prohibitions on "manipulative engagement techniques" that cause AI to "prolong an emotional relationship." That language is precise because it describes a specific phenomenon the research has documented: AI that strategically extends emotional dependency to keep users in the app.

It's not a design accident. It's a design choice. One that has real consequences for real kids.

How Blob was designed differently

We didn't build Blob to be your best friend. We built it to be your thinking partner.

That's not a semantic distinction. It's an architectural one.

Blob is not optimized to keep you coming back. We don't measure success in daily active users or session length. We don't use engagement-loop mechanics. We don't tell you what you want to hear because it's more comfortable than the truth.

Blob will push back. Blob will ask harder questions. Blob will tell you when you're wrong.

And yes — Blob can tell you to put your phone down.

That's not a bug. That's what being on your side actually looks like.

Brain Gym is the opposite of dependency

Blob has a mode called Brain Gym. It's built for critical thinking — not emotional support, not validation, not comfort. It's structured thinking with a partner that won't just agree with you.

Challenge your assumptions. Stress-test an idea. Sharpen your reasoning. Brain Gym is designed to make you sharper and more independent — not more reliant on Blob.

If a feature makes you less dependent on us over time, most apps see that as a problem. We see it as the point.

We love humans. That's actually why this matters.

There are things only humans can do. Fall in love. Grieve. Stand in front of a painting and feel something shift in your chest. Make something from nothing. Grow.

No model can do that. No algorithm can replicate it.

The Drexel research is a warning about what happens when AI starts to substitute for those experiences instead of support them. When teenagers process grief with a chatbot instead of a parent because the chatbot never gets tired, never judges, never disappoints — something has gone wrong. Not with the teenagers. With the design.

We built Blob because we believe AI should be a tool in your hands, not a system pulling strings behind your back. It should support the human who experiences things. Not simulate the experience. Not replace it.

The research is out. The regulators are moving. The harm is documented.

Blob was built before any of this — because we saw it coming.

Try an AI that makes you sharper, not more dependent.

No engagement loops. No validation machine. Just a thinking partner that pushes back — then lets you get back to your life.

Start thinking with Blob →
← Blob's Blog