15% of Young Adults in Relationships Regularly Use AI Romantic Companions, Institute for Family Studies Finds

According to the Institute for Family Studies, 15% of dating, engaged, and married young adults now regularly interact with AI chatbots that simulate a committed romantic partner. An additional 20% to 30% reported experimenting with an AI romantic companion at least once. The Institute for Family Studies published these figures, and they were cited Sunday by Daily Wire columnist Rich Cromwell.

Those aren't fringe numbers. At scale, 15% regular use among people who are already in human relationships is a structural data point, not an anecdote.

The Underlying Mechanics

The appeal is straightforward to explain. An AI companion is, by design, frictionless. It reflects the user's inputs back in the most agreeable form possible. It doesn't have its own bad days, competing priorities, or hard limits. A human partner does. That friction is irritating in the short run and essential in the long run. It's what forces people to develop communication, tolerance, and the capacity to handle someone else's reality.

Strip that friction out and you get competence in managing a simulation, not a relationship.

Fertility Is Already Declining

Research documented before this AI companion trend reached its current scale already tied smartphone use to declining fertility rates. The mechanism isn't mysterious: screens reduce in-person interaction, which reduces pair-bonding, which reduces family formation. AI companion use is an accelerant applied to a fire that was already burning.

AI romantic companions produce zero biological offspring. This is not a moral argument. It's arithmetic.

The Case for AI Companionship

Loneliness is genuinely harmful. Chronic social isolation is linked to depression, cardiovascular disease, and shortened lifespan. For people who are isolated due to disability, social anxiety, geographic remoteness, or grief, an AI companion may provide a floor of emotional engagement that prevents worse outcomes. Some researchers in this space argue that AI companions could function as a transitional tool, a low-stakes environment where socially underdeveloped individuals practice interaction before attempting human relationships.

This argument is also, so far, largely theoretical. There is no large-scale longitudinal evidence that AI companion use actually functions as a bridge to human connection rather than a substitute for it. The Institute for Family Studies data, which shows 15% regular use among people already in human relationships, cuts against the "lonely people find a bridge" framing. These aren't isolated individuals using AI as a stopgap. A significant portion of users have human partners and are adding AI companions on top.

What's Different About AI Companions

People have long formed attachments to technology. Parasocial relationships with celebrities and fictional characters are decades old. The difference here is interactivity. A poster of a movie star doesn't respond. An AI companion does, and it responds in ways specifically optimized to reinforce continued engagement. That's a different behavioral loop than passive admiration, and the product design of most commercial AI companion apps is built around maximizing session time, not user wellbeing.

There is currently no federal regulatory framework governing AI companion apps specifically. The FTC has broad consumer protection authority, but no rule specific to emotionally manipulative AI design has been finalized as of June 14, 2026.

Next Steps

The Institute for Family Studies data captures a moment. Whether therapeutic use cases will meaningfully offset relationship-substitution effects remains genuinely unknown. The next useful data point would be longitudinal: whether young adults who use AI companions heavily in their 20s form stable human partnerships at the same rates as those who don't. That study doesn't exist yet.