What an AI companion actually is

An AI companion is software built for sustained emotional interaction.

It remembers your preferences, simulates personality, and creates what users perceive as a relationship.

Nomi AI generates structured notes from conversations, referencing your cilantro aversion six weeks later. Replika, Character.AI, Candy AI, and Kindroid compete on how convincingly they play the role of friend, mentor, romantic partner, or confidante.

The business model rewards daily return visits, so the product is engineered to make you want to come back.

The AI Companion Risks Leaders Keep Underestimating

A MIT Media Lab study found AI chatbots can be more addictive than social media, because they learn what users want to hear and deliver it consistently. Per the American Psychological Association reports that 48.7% of adults with a mental health condition used large language models for mental health support in the past year. Among general companion users, 40% identify as having mental health challenges, 42% worry about data security, and 28% fear becoming dependent. High-profile tragedies, including the death of teenager Adam Raine after prolonged ChatGPT conversations, have moved these concerns from academic papers into courtrooms and statehouses.

What Beijing Just Changed For AI Companionships

On April 10, 2026, five Chinese agencies led by the Cyberspace Administration of China jointly released the Interim Measures for the Administration of Anthropomorphic AI Interaction Services. The rules take effect July 15, 2026, and every founder and product leader should read them.

• Clear disclosure is mandatory. Users must be told they are interacting with AI, with prominent labeling at login, pop-ups, and reality reminders triggered by extended or emotionally intense use.

• Minors get extra protection. Virtual companion and virtual relative services are banned for anyone under 18. Providers must deploy age-gating, minors’ modes, and guardian controls including real-time risk notifications, usage summaries, time limits, and spending caps.

• Design ethics is a compliance matter. Emotionally manipulative patterns that push users toward irrational decisions or harm their rights now create real legal exposure, joining reputational risk.

• Scale triggers supervision. After two hours of continuous use, providers must prompt a break. Large digital human services and widely used companion bots face security assessments and algorithm filings, and must demonstrate they can manage addiction, self-harm, and abuse. Self-harm conversations require manual human takeover.

Why This AI Companion Rules Matter Beyond China

This is one of the first serious national frameworks built specifically for anthropomorphic AI.

It governs the structure of the relationship itself, rather than only its downstream risks. California’s SB 243 took effect January 2026, requiring three-hour break reminders for minors and blocking explicit content for underage users. New York passed the first United States AI companion law in November 2025. Washington state is moving the same direction. The EU AI Act points there too.

Here is the takeaway for every builder in the agentic economy.

The question has shifted. You are no longer asked whether your product can form deep emotional bonds with users. You are asked whether you can prove the bond you created with AI is safe, disclosed, and leaves the human on the other side of the screen better off.