AI Chatbots Are Dating's New Normal. Platforms Are Letting It Happen.

We've crossed a threshold in online dating: the default assumption is no longer that you're talking to a person. One in four American singles now deploy AI to write their messages, whilst the platforms profit from the engagement boost and treat the resulting authenticity crisis as an acceptable externality. The industry's response to this fundamental erosion of trust has been a collective shrug.

Match Group CEO Spencer Rascoff told Bloomberg that off-platform AI use is 'difficult to control'. Hinge CEO Jackie Jantos offered the usual platitudes about human connection whilst declining to implement any actual restrictions. Dating platforms universally ban AI-manipulated photos—deepfakes, face-swaps, the usual suspects—but explicitly permit AI-written text. Visual deception: prohibited. Conversational deception: crack on.

Create a free account

Unlock unlimited access and get the weekly briefing delivered to your inbox.

Register free

No spam. No password. We'll send a one-time link to confirm your email.

The authenticity crisis that drove members to niche platforms and off-app connections is about to get considerably worse. Trust & safety teams should be treating this as a Code Red event. They're treating it as an acceptable externality.

The double standard

Here's what's consistent: platforms will invest heavily in verification tools for anything they can measure and control. Photo verification rolled out industry-wide precisely because it's technically straightforward—submit a selfie, match it to your profile, done. Bumble even added ID verification. The rationale is sound: visual impersonation creates immediate safety risks, undermines member trust, and generates negative press.

AI-generated text presents the inverse problem. There's no reliable technical solution for detecting it at scale. Large language models produce output indistinguishable from human writing—that's rather the point of them. The platforms could ban the practice outright and enforce it through terms of service violations when reported. They could implement friction points: rate limiting on message volume, mandatory pauses before replies, in-app warnings about AI use. They're doing none of this.

The reason isn't technical capability. It's incentive alignment. AI-assisted messaging increases reply rates, extends conversation threads, and boosts session duration—all metrics that dating apps optimise for relentlessly. A user who deploys ChatGPT to craft the perfect opening line is more likely to get a response than one agonising over writer's block. More responses mean more engagement. More engagement means better unit economics.

What gets sacrificed is the thing dating apps claim to provide: genuine human connection.

The ecosystem taking root

The market has responded with characteristic speed. Rizz, Winggg, and YourMove AI have built entire businesses around dating message generation. Users upload screenshots of conversations, receive suggested replies, paste them back into Tinder or Hinge. Some tools analyse conversation history to identify why chats stall. Others generate opener suggestions based on match profiles.

Rizz's 1.5 million monthly active users—disclosed in 2024—represents meaningful scale. For context, Bumble reported 3.8 million paying subscribers in Q3 2024. A standalone AI dating assistant has achieved user numbers equivalent to 40 per cent of Bumble's paying base. That's not a fringe behaviour. That's a parallel infrastructure.

These tools operate entirely off-platform, which conveniently absolves dating apps of any responsibility for their proliferation. Rascoff's comment about difficulty controlling off-platform AI use is technically accurate. It's also a convenient abdication. Platforms control their terms of service. They control what behaviours result in account suspension. They've chosen not to make AI-assisted messaging a violation.

The Norton figure—60 per cent of daters believing they've encountered AI messages—deserves scrutiny. This is perception, not verified detection. But perception shapes behaviour. When a majority of users assume smooth, well-crafted messages are chatbot output, the incentive to invest effort in genuine communication collapses. Why spend twenty minutes crafting a thoughtful opener when your match assumes you used ChatGPT anyway? The equilibrium tips toward automation.

Platform burnout made this inevitable

The industry would prefer to frame AI assistance as user innovation—singles finding creative solutions to improve their experience. The less comfortable framing: this is a symptom of profound product failure. Swipe-based matching has created an environment so exhausting, so depleting, that users are paying third-party services to handle basic conversation.

Multiple surveys—tracked in DII's quarterly sentiment analysis—show sustained increases in dating app fatigue. Users report spending hours on apps without meaningful connections, match volumes that overwhelm rather than enable, and conversations that go nowhere. The business model works: time on platform drives ad impressions and conversion to paid tiers. The user experience fails: that time feels wasted.

AI messaging tools are a rational response to irrational conditions. The problem isn't writer's block. The problem is that the current model requires users to conduct high-volume, low-investment conversations with strangers, and humans find that task soul-destroying.

If the platforms provided matching that actually worked—smaller match volumes with higher compatibility, conversation prompts that generated genuine dialogue, friction that filtered for intent—users wouldn't need Claude to write their replies.

What happens when everyone's a chatbot

The trajectory is straightforward. As AI messaging becomes standard practice, arms race dynamics take hold. Users who don't deploy AI tools find themselves at a conversational disadvantage. Reply rates for human-written messages decline relative to AI-optimised ones. The baseline expectation for message quality rises—what used to read as thoughtful now reads as minimum viable effort.

Platforms face a legitimacy crisis. Online dating's value proposition rests on facilitating genuine human connection. When the early stages of that connection are algorithmically mediated, the promise collapses. We're not far from a scenario where two chatbots carry out the initial conversation, notify their respective humans when sufficient rapport has been established, and hand off to a first date between two people who've never actually spoken.

The industry's silence on regulation suggests operators believe this is manageable. They're wrong. Trust & safety teams should be modelling the knock-on effects: increased reports of catfishing when AI-assisted personas don't match in-person behaviour, member churn as authenticity concerns drive users to alternative platforms or off apps entirely, and regulatory attention once enough users complain to legislators.

Rascoff's comment about difficulty controlling off-platform AI might be true for now. But the EU Digital Services Act (DSA) and UK Online Safety Act (OSA) both require platforms to assess and mitigate risks from foreseeable misuse. 'We can't detect it' won't satisfy regulators for long, particularly when the platforms are actively choosing not to prohibit it. Compliance teams should be gaming out how to respond when the ICO or Ofcom starts asking questions.

The obvious move—banning AI-generated messages outright and enforcing it through reporting mechanisms and pattern detection—would require platforms to accept lower engagement metrics in exchange for preserving trust. That's a trade Match Group, Bumble (BMBL), and the rest have shown no willingness to make. Which means the authenticity crisis will deepen until it manifests in churn rates significant enough to show up in earnings calls. By then, the damage will be structural.