47% of US Singles Reject AI in Dating Apps

Nearly half of US singles—47%, according to Match Group's own data—hold a negative view of artificial intelligence in dating applications. The figure appears in a report published this week by TechCrunch and comes from internal surveys by the company that operates Tinder, Match.com, and Hinge, among other platforms. It's a significant number: product designers in dating have been betting on integrating conversational assistants into their apps for months, and this percentage calls into question how quickly that bet should move forward.

What's interesting is that the rejection isn't uniform. The same study shows that a meaningful portion of those users—including some who say they're skeptical—do accept AI when its function is concrete and limited: polishing a profile description or suggesting an opening message. The line many people draw, it seems, isn't "I don't want AI here," but rather "I don't want AI to replace real human interaction."

What the report actually says

Match Group hasn't published complete methodology, so the percentages should be read with some caution. What does emerge from the summary is a fairly clear tension:

• General rejection: 47% perceive AI negatively in the dating context.
• Functional acceptance: a portion of users is willing to use AI for low emotional-risk tasks—optimizing a profile, breaking the ice in a chat—as long as the other person doesn't know a model is involved.
• Distrust of authenticity: the main fear doesn't seem to be privacy, but rather not knowing whether the person writing is real or an automated agent.

This last point is most relevant for any product team. Dating apps essentially sell the promise of human connection. Introducing AI without active transparency triggers exactly the fear that most damages that promise.

Why it matters beyond dating

The dating sector is, in many ways, an accelerated laboratory for how users react to AI in personal contexts with high emotional stakes. What happens on Tinder or Hinge tends to foreshadow attitudes that later reappear in other categories: mental health assistants, productivity coaches, HR chatbots.

The pattern that emerges here—acceptance of AI as a support tool, rejection of AI as a substitute for a person—isn't new, but it's rarely quantified with data at this scale. Match Group has access to tens of millions of active users, which gives some statistical weight to the figure, though the lack of methodological detail prevents treating it as an academic reference.

For those integrating language models into consumer products—whether with Claude, proprietary APIs, or any other stack—the finding suggests something practical: disclosure isn't just an ethical matter, but also one of retention. Users who discover they've been chatting with an agent without knowing rarely stick around.

Who should care about this data

The report is particularly relevant to three groups:

1. Product managers of apps with social components who are evaluating how much automation to introduce into conversation flows.
2. Conversational agent design teams that need to calibrate how far an assistant can go without creating friction.
3. Investors and analysts in consumer AI, who have been modeling adoption rates for months without hard data on resistance.

What the Match Group study doesn't answer—and would be interesting to see—is whether rejection varies by age group, by previous experience with AI tools, or by the specific type of integration. Those breakdowns would shift the reading considerably.

---

Our take is that this kind of data should temper the enthusiasm with which some product teams are integrating generative AI into flows where the perception of authenticity is the core asset. Measuring resistance before scaling isn't pessimism; it's sound engineering.