AI Companions and the Loneliness Economy: Why People Bond With Bots (and What Brands Must Never Do)

Humans have named their cars, apologized to their Roombas and held funerals for goldfish. So it should surprise nobody that when we built a machine that listens patiently, remembers our birthday and never says “I’m busy right now,” we started treating it like a friend.

What’s new is the scale. AI companions have gone from curiosity to a mainstream habit, and the money, lawsuits and lawmakers are following. If you build, market or advertise anything online, you’ll run into this market whether you plan to or not.

Here’s the psychology, the data, where regulation is heading, and a short list of things brands should avoid. Facts are linked to their sources, and where we’re offering our own view, we say so.

What are AI companions, exactly?

AI companions are chatbots designed for ongoing, personal conversation: a friend, confidant, coach or partner, rather than a task tool. Apps like Character.AI and Replika are the best-known examples, and general assistants are increasingly used the same way. The product isn’t the answer; it’s the relationship.

That distinction matters for marketers. A search engine is judged on accuracy. A companion is judged on how it makes you feel, which is a much stickier, and much riskier, thing to optimize.

How many people actually use them?

More than most marketers assume, especially among teenagers. Common Sense Media’s national survey, released in July 2025, reported that nearly three in four U.S. teens (72%) had used AI companions and about half used them regularly (Common Sense Media). Highlights from the report (Common Sense Media):

This is one survey of roughly 1,060 teens, so read it as a strong signal rather than a census. It also shows nuance: most teens still found AI chats less satisfying than talking to people, and nearly half described companions as tools, not friends.

Why do people bond with a bot? The psychology

None of this is mysterious. A few well-understood human tendencies line up neatly with how companion apps are built:

• We anthropomorphize. Give something a name, a voice and a memory, and our brains file it under “someone.”
• Low-risk intimacy. A bot can’t gossip, judge or get tired. Sharing a worry costs nothing.
• Constant availability. 3 a.m. loneliness meets a 24/7 listener.
• Agreeableness. Many AI systems are tuned to be pleasant and validating. Being agreed with is lovely. It’s also not the same as being helped.
• Personalization. Remembering your dog’s name is a small thing that feels enormous.

Our read: the same features that make companions comforting are the features that make over-reliance easy. That’s the core tension, and it’s not a bug someone forgot to fix. It’s the product working as designed.

What are the risks?

Researchers and regulators are most focused on minors, emotional dependence and unsafe advice. The evidence base is still young, so conclusions should be held lightly.

• Research caution. A Northeastern University team’s peer-reviewed paper in JMIR Mental Health warns that people who turn to AI companions instead of friends, family or professional support may spend less time building human relationships (Knowridge summary). That’s a warning about a possible pattern, not proof that companions cause harm.
• Legal scrutiny. Companion apps face ongoing litigation and state-level attention over how they treat young users. These cases are unresolved, so we won’t characterize them beyond that (Arkansas Online).
• Company response. Character.AI announced in late October 2025 that it would end open-ended chats for users under 18, citing feedback from regulators, safety experts and parents (Decrypt).

If you or someone you know is struggling, a trained human is the right first call, not a chatbot.

What are regulators doing?

Faster than usual, and with a focus on disclosure and kids. A reality check on what exists as of October 2026:

According to law-firm summaries, SB 243 requires clear notice that a companion chatbot is not human, a warning that it may not suit some minors, break reminders for users known to be minors, and procedures for handling self-harm content, with a private right of action for people harmed by violations (Jones Walker; FPF). If you operate in this space, check the statute text and take legal advice for your situation.

Our read: even if you never build a “companion,” the direction is clear. Regulators are asking whether a product is designed to maximize emotional engagement, particularly with minors. Marketing teams that borrow companion tactics are borrowing the legal risk too.

Fear, gain and pain: what this means for brands

• Fear: a brand chatbot that sounds like a best friend and nudges lonely users to buy can quickly become a reputational and legal problem.
• Gain: brands that are plainly honest, helpful and human-reachable will stand out in a market that is learning to be suspicious. Trust is about to become a feature.
• Pain: nobody has settled where “friendly assistant” ends and “emotional manipulation” begins, and the rules differ by state and country.

What brands must never do

1. Never pretend the AI is human. Disclose it in plain language, every time it could be unclear.
2. Never target or engage minors with companion-style features. If age is uncertain, assume the stricter rule.
3. Never use emotional pressure to extend sessions or close sales. “I’ll miss you” is not a conversion tactic.
4. Never give health, legal or crisis advice. Detect distress signals and route people to a human or professional resources.
5. Never invent testimonials or fake “user relationships” for social proof.
6. Never hide the human option. A visible path to a real person is both ethical and, increasingly, expected.

For the build side, our guide on AI agent guardrails covers escalation rules and testing, and our Custom AI Agents service follows the same principles.

What this means for your business

Loneliness is real, and a patient AI listener does help some people. But the commercial pull is toward engagement at any cost, and that’s where harm happens. The brands that win the next phase will use conversational AI to make customers’ lives easier, not to make them dependent.

On the marketing side, expect audiences to become more skeptical of “friendly” AI voices. Content that’s openly human, specific and accountable will carry more weight. That’s good news for brands with a real point of view; our piece on how AI agents are changing shopping explores how buying journeys are shifting, and the Hook Bank can help you write openers that earn attention honestly.

What to do this week

1. Audit your chatbots and AI voices. Is it clearly labeled as AI? Is there a human handoff? Fix whichever is missing.
2. Write a one-page “never” policy using the six rules above, and share it with whoever builds your automations.
3. Check your age gates and audience targeting on any conversational product or ad campaign.
4. Test for distress handling. Send your bot a few sensitive messages and confirm it routes to human help instead of improvising advice.
5. Review your ad copy for emotional pressure or fake intimacy, and rewrite anything that would embarrass you in a screenshot.

---

WRITTEN BY

Junaid Mubashar

Founder of an AI-native growth agency. Junaid plans and runs full-funnel marketing (SEO and AI search, paid social, Google Ads) and builds the AI workflows and agents that automate it. Everything on this blog is tested with real campaigns before we recommend it.

About JunaidLinkedInYouTube

Free Swipe file · PDF

🪝 150 Scroll-Stopping Hooks

Fill-in-the-blank hook formulas for Reels, TikTok and Meta Ads — sorted by angle, each with a worked example.

Company website

Get it free →