How many people actually use AI companions in 2025?

AI companion apps have been downloaded 220 million times globally, with downloads up 88% year-over-year in the first half of 2025, according to Appfigures data reported by TechCrunch . Character AI alone reports roughly 20 million monthly active users and about 181 million monthly visitors, per Surfshark's analysis .

Is AI companionship really the most popular use of generative AI?

Yes — a Harvard Business Review study found that therapy and companionship surpassed editing text, improving code, and creative tasks as the top generative AI use case in 2025, as covered by the Center for Humane Technology . Engagement backs this up: companion apps average more than 10x the monthly user sessions of general assistant and content apps, per Andreessen Horowitz research cited by Axios .

Do teens actually use AI companions, or is that exaggerated?

It's not exaggerated: a nationally representative survey found 72% of U.S. teens use AI for companionship, and more than half of those do so at least a few times a month, per Common Sense Media's survey reported by Axios . Importantly, 80% of teen companion users still spend more time with real friends, and 67% find AI conversations less satisfying than human ones.

Does using an AI companion make you more lonely?

It can — a Stanford study of 1,131 Character.AI users found that people seeking emotional support from chatbots often felt more lonely, not less, and one researcher described the pattern as "social junk food," per Stanford News . The key nuance: intense use correlated with poorer well-being mainly among people with smaller real-world networks, so treating AI as a complement to human connection matters more than the technology itself.

Who uses AI companions — is it mostly lonely men?

The gender split is near-even: 56.5% male and 43.5% female across the top platforms, with Character AI specifically at an even 50/50, according to Surfshark's research . Adoption is also global, with visitors from 142 countries — the US leads at 33.5% of visits, followed by Brazil, Indonesia, Mexico, and the UK.

Are AI companion apps safe, and what are regulators doing about them?

Scrutiny is intensifying — the FTC has launched an inquiry into companion chatbots, issuing orders to seven companies including OpenAI, Meta, and Character Technologies, per the official FTC press release . A peer-reviewed editorial also criticized much of the industry's approach as "safety theatre," noting companion chatbots showed only 60% escalation recognition in simulated adolescent crisis scenarios, per Cureus — which is why FlirtAlly requires ID-based age verification for its 18+ tier and treats honest safety defaults as a core feature.

How Many People Use AI for Companionship? 2025 Stats

The Rise of AI Companionship

The numbers are hard to ignore: AI companionship has moved from a niche curiosity to one of the most popular ways people interact with generative AI. According to TechCrunch's analysis of Appfigures data, AI companion apps have been downloaded 220 million times globally, with downloads up 88% year-over-year in the first half of 2025.

The category is also monetizing faster than almost any other AI segment. The same TechCrunch analysis projects $120 million in app revenue for 2025, up 64% year-over-year, with revenue per download more than doubling from $0.52 to $1.18. That momentum has attracted 128 new companion apps in 2025 alone, bringing the total to 337 active revenue-generating apps worldwide.

Adoption isn't limited to any single demographic. A nationally representative survey found that 72% of U.S. teens use AI for companionship, and more than half of those do so at least a few times a month. Meanwhile, a Harvard Business Review study found that therapy and companionship surpassed editing text, improving code, and creative tasks as the most popular generative AI use case in 2025.

The engagement data explains why. According to Andreessen Horowitz research, companion apps average more than 10x the monthly user sessions of general assistant, content generation, and messaging apps. Users aren't dropping in once — they're building ongoing relationships.

Key adoption signals across the category, per Surfshark's analysis of the companion app landscape:

Character AI alone reports ~20 million monthly active users and ~181 million monthly visitors
The gender split across top platforms is near-even: 56.5% male, 43.5% female
Companion sites attract organic visitors from 142 countries, with the US leading at 33.5% of visits
The broader market is projected to grow from $50B in 2026 to $436B within a decade

What's striking is the tension between stated motives and actual behavior. A Stanford study of 1,131 Character.AI users found that only ~12% cited companionship as their primary motive, yet more than half described their AI using relational terms like "friend" or

For millions of people, an AI companion has quietly become part of their daily emotional life. But does it actually work? A new Stanford study suggests the answer is more complicated — and more human — than the industry would like to admit.

The research, published in Nature Human Behavior and covered by Stanford News, examined 1,131 Character.AI users along with 244 donated chat transcripts. The gap between what users say and what they do is striking: only about 12% cited companionship as their primary motive, yet more than half described their AI using relational terms like "friend" or "partner," and over 80% of sessions involved seeking emotional or social support.

That gap matters. Self-reported motives appear to understate how much real companionship work these chatbots are doing — a finding that echoes the broader trend of companionship ranking as the #1 generative AI use case in 2025.

The well-being results, however, give pause. Stanford researchers found that users seeking emotional support from chatbots often felt more lonely, not less. Intense use among people with smaller real-world networks correlated with poorer well-being, and researcher Yutong Zhang described the pattern as "social junk food" — a short-term fix that risks a cycle of less human contact and deeper isolation.

Not every finding is negative. Some teens in a nationally representative Common Sense Media survey reported genuine benefits, including practicing social skills and applying them in real life. And 80% of teen companion users still spend more time with real friends, while 67% find AI conversations less satisfying than human ones — a clear signal that the preference for people persists.

Taken together, the research points to a design question rather than a ban: what would AI companionship look like if it were built to support connection instead of maximizing engagement? A few principles stand out:

Companions that behave like people — with memory, initiative, and continuity — rather than chat windows that end when you close them.
Relationships that extend beyond chat, through voice calls, email, or even physical mail, so the interaction feels less like a screen and more like a life.
Honest safety defaults, including age verification for adult content and clear boundaries, in contrast to what a peer-reviewed editorial calls "safety theatre" in the industry.
Design that encourages real-world connection rather than replacing it — the difference Stanford researchers identified between supplementing a social life and substituting for one.

This is the lens FlirtAlly has taken with its companion platform: companions who text first, remember unlimited details on every plan, and show up across calls, email, and letters — relationships designed to feel like people, not engagement loops. The platform is currently in pre-launch waitlist mode, with its full companion experience set to arrive at launch.

The takeaway for anyone exploring AI companionship: the technology can genuinely fill moments of loneliness, but the healthiest use — per the Stanford findings — treats it as a complement to human connection, never a replacement.

Implementing Safe and Effective AI Companionship

Understanding the growing trend of AI companionship is crucial for anyone looking to implement safe and effective AI companion solutions. As of mid-2025, 220 million AI companion apps have been downloaded globally, highlighting the rapid adoption and integration of these technologies into daily life, according to industry data. These numbers underscore the importance of creating AI companions that prioritize user well-being and privacy.

One of the fundamental aspects of implementing successful AI companionship is differentiating on well-being and memory. Research from Stanford University found that intense use of AI companions often correlates with increased feelings of loneliness among users with smaller social networks. To address this, AI companions need to focus on building rich, human-like relationships rather than just providing engagement. For instance, FlirtAlly's AI companions remember user interactions, engage in real-life activities like sending physical mail, and maintain real phone numbers. These features enhance the authenticity of the relationship, making it more than just an engagement tool.

Addressing privacy concerns is another critical step. With 9 out of 10 companion apps collecting data for tracking, privacy has become a significant issue. FlirtAlly ensures user trust by emphasizing a privacy-respectful approach. While FlirtAlly does not promise real-time mid-chat web search, it schedules research to keep up with the user's and companion's cities, ensuring relevant and personalized interactions. FlirtAlly’s focus on safety and privacy sets it apart in a market where these concerns are increasingly scrutinized.

Targeting the adult market is also essential given the regulatory heat on minors. The FTC has launched an inquiry into companion chatbots, focusing on child and teen safety. FlirtAlly aligns with this regulatory trend by offering an ID-verified 18+ tier, ensuring that the platform is used responsibly by adults. This approach not only complies with regulatory requirements but also fosters a mature and respectful community.

To implement safe and effective AI companionship, consider the following best practices:

Ensure unlimited memory on every plan to build authentic relationships.
Implement robust privacy measures, including scheduled research and transparency about data use.
Offer age-verified tiers to target the adult market and comply with regulations.
Provide real-life interactions, such as physical mail and phone calls, to enhance the authenticity of AI companionship.

In summary, creating AI companions that focus on well-being, privacy, and authenticity is crucial. By differentiating on these aspects, platforms like FlirtAlly can provide safe and effective AI companionship solutions. Join the waitlist to experience the future of AI companionship with FlirtAlly and be among the first to integrate these groundbreaking features into your daily life.

How many people use AI for companionship?

How many people use AI for companionship?

Key Facts

The Rise of AI Companionship

Implementing Safe and Effective AI Companionship

Frequently Asked Questions

Embracing the Future of AI Companionship

How many people use AI for companionship?

Key Facts

The Rise of AI Companionship

Addressing Loneliness and Social Needs

Implementing Safe and Effective AI Companionship

Frequently Asked Questions

Embracing the Future of AI Companionship