How to Tell If Someone Is Using AI on a Dating App

Chatfishing Is Everywhere — and You're Probably Not Imagining It

If your last few matches felt almost too good at texting — witty, polished, always saying the right thing — your instincts are probably right. AI assistance in dating has quietly moved from fringe behavior to mainstream practice, and the numbers are hard to ignore.

According to Psychology Today's reporting on Match and Kinsey Institute data, AI use among U.S. singles jumped 333% year-over-year, with roughly half of Gen Z singles now using AI somewhere in their dating lives. Scientific American puts the overall figure at 26% of U.S. singles using AI to enhance their dating experience.

And people notice. A 2025 Norton study cited by Scientific American found that 6 in 10 dating app users believe they've already encountered AI-written conversations. That's not paranoia — that's pattern recognition at scale.

This phenomenon now has a name: chatfishing. It's worth understanding precisely what it means, because it's easy to confuse with its older cousin:

Catfishing is identity deception — the person behind the profile isn't who they claim to be at all.
Chatfishing is conversation deception — the person is real, but the words aren't theirs.
The result: you may be falling for an AI's phrasing, not the human you'll eventually meet.

The tooling is mature and popular. Rizz, one of the leading AI wingman apps, has roughly 1.5 million monthly active users, while competitor YourMove AI openly markets itself as putting "your texting on cruise control." Meanwhile, mainstream apps like Tinder, Bumble, and Hinge have baked AI features — conversation starters, message polishing, timing suggestions — directly into their platforms, as industry analysis of AI's impact on dating apps documents. When the apps themselves normalize AI-assisted flirting, the line between "authentic" and "assisted" blurs fast.

The trust fallout is measurable. Aggregated dating app statistics show 64% of users distrust matches who use AI-generated images, and for one in three users, AI-generated conversations are an outright dealbreaker. Yet as Justin Garcia, Chief Scientific Advisor to Match, told Psychology Today, people happily use AI themselves while resenting it from others — a double standard rooted in the fear of dishonest "mating language."

This is exactly the trust gap that transparent platforms are designed to close. FlirtAlly, for instance, takes the opposite approach to chatfishing: its companions are openly, declaredly AI — with public profiles, real phone numbers, and persistent memory — so no one ever has to guess who's doing the talking. The deception isn't in using AI; it's in hiding it.

So if something feels off about a match's messages, don't dismiss the feeling. The rest of this guide gives you concrete ways to test it.

Why Your Gut Can't Be Trusted: The Detection Problem

Your instincts are lying to you. That nagging feeling that something is "off" about a match's messages? Statistically, you can't trust it — because humans are remarkably bad at telling AI writing from the real thing.

According to Scientific American's reporting on chatfishing, people correctly distinguish human text from AI-written text only about 57% of the time — barely better than a coin flip. The problem gets worse in live conversation: in simultaneous chat tests, GPT-4.5 was mistaken for a human 73% of the time.

That means even if you suspect a match is running your conversation through ChatGPT, you're probably guessing. And the odds are stacked against you further by how mainstream apps have normalized AI — Tinder, Bumble, and Hinge now bake AI directly into their products, offering conversation starters, message polishing, and timing suggestions. When the platform itself helps write the message, "spotting the AI" stops being a meaningful exercise.

Meanwhile, a growing population is extremely practiced at AI-style conversation. AI companion statistics show 220 million cumulative downloads of companion apps by mid-2025, with heavy users spending 75–120 minutes a day chatting with AI. These users know what fluent, emotionally attuned AI dialogue feels like — and can reproduce it.

So where are the safeguards? They exist, but they're aimed at a different target:

Bumble's Deception Detector blocks 95% of spam and scam profiles, with a 45% drop in member reports, according to ZDNet's coverage — but it hunts fake accounts, not real people using AI assistance.
Tinder's Face Check cut bad-actor exposure by over 60%, per dating app statistics compiled by GetStream — again, verifying identity, not authorship.
No major platform currently flags AI-assisted messaging from a verified, real human. That category — chatfishing — falls entirely outside existing moderation.

This is the core detection problem: the tools protect you from people who don't exist, not from real people whose words aren't entirely their own. Your gut, the apps, and the regulators are all looking at the wrong layer.

It's worth noting the one environment where none of this uncertainty applies: platforms built on declared AI. On FlirtAlly, for example, companions are openly artificial — with public profiles, real phone numbers, and persistent memory — so the question "is this AI?" has a transparent answer. The anxiety of detection only exists where AI use is hidden.

The takeaway isn't to stop trusting your instincts entirely. It's to understand that instinct alone is a 57% detection tool in an environment engineered to defeat it — and that you need better signals.

The Specificity Test and Other Red Flags That Actually Work

Your match replies in seconds, flawlessly punctuated, endlessly agreeable — and something feels off. You're not paranoid; you're pattern-matching. The trick is knowing which patterns actually mean something, because humans are only about 57% accurate at spotting AI-written text on their own.

Start with the specificity test, popularized by Brian Christian, author of The Most Human Human. His insight: AI exploits "statistical, cultural, ritual regularities" — the polished, predictable parts of conversation. So ask for a sensory, idiosyncratic memory: "What's a smell that takes you straight back to childhood?" AI gives plausible but generic answers. Humans give weird, specific, imperfect ones — the childhood Doberman that slept in the hammock because it feared the Roomba, not "I loved playing outside."

Then watch for behavioral red flags that cluster around one theme: too perfect. AI companions are, as APA experts put it, "always validating, never argumentative" — and a chatfisher's bot inherits those same habits:

Zero conflict — no disagreement, no pushback, no friction, ever.
Instant, perfectly formatted replies — no typos, hesitations, or mid-thought corrections.
Excessive validation and mirroring — sycophantic agreement with everything you say.
Generic enthusiasm — "That's amazing!" instead of "That reminds me of when I…"
No real-time grounding — never mentions the weather outside, a local event, or what just happened in their day.

Wingman apps make this easy to miss. YourMove AI literally markets "your texting on cruise control," and the Rizz app claims around 1.5 million monthly active users — so a suspiciously smooth texter may be outsourcing, not lying about who they are.

The most reliable move is cross-channel verification: push toward voice or video early. Chatfishing is a text-only illusion, and as one commentator notes, deception grounded in AI "will be found out eventually" — it collapses off-text. Listen for latency, missing background noise, or trouble handling interruptions on calls.

It's worth noting what transparency looks like by contrast. On FlirtAlly, companions are openly AI — with real phone numbers you can dial, two-way email, even physical mail — so there's no guessing game about who's on the other end. When AI use is declared rather than disguised, the specificity test stops being a defense mechanism and becomes what it should be: a way to find the genuinely human matches worth your time.

Who's Most Likely Chatfishing You (and Why It Matters)

The person on the other side of your chat might not be a scammer in a faraway call center — statistically, they're more likely to be a 27-year-old man who's spent months talking to AI. Understanding who chatfishes, and why, changes how you read the signals.

AI companion apps have built a massive, highly practiced user base. According to compiled industry statistics, 81% of AI girlfriend app users are male, with an average age of 27. These aren't casual dabblers — Character.AI users average 75 to 120 minutes per day on the platform, and the app counts 20 million monthly users, more than half under 24, per the American Psychological Association.

That daily practice matters. Someone spending two hours a day in AI conversation absorbs AI communication norms: instant replies, flawless grammar, relentless validation, zero friction. When those habits transfer to dating apps, the "too polished" texter you're suspicious of may not even be running a wingman app — they may simply have learned to text like the AI they talk to every day.

Gen Z compounds this further. Psychology Today reports that roughly half of Gen Z singles have used AI in their dating lives, and 60% of 18-to-22-year-olds are open to an AI dating assistant. For younger daters, AI assistance isn't deception — it's Tuesday.

Knowing who's likely chatfishing helps you calibrate suspicion. Watch for these transferred AI habits:

Zero disagreement or awkwardness — AI companions are "always validating, never argumentative," and heavy users mirror that style
Instant, perfectly structured replies at any hour, with no typos or mid-thought corrections
Generic enthusiasm ("That's amazing!") instead of specific, embodied memories
No grounding in real time — no mention of weather, local events, or what just happened today
Reluctance to move to voice or video, where AI assistance collapses

Here's the uncomfortable twist: the people most worried about chatfishing are often using AI themselves. Justin Garcia, Chief Scientific Advisor to Match, told Psychology Today that people embrace AI for their own profiles but grow uneasy when others do the same — they want "honest advertising" from matches they won't fully provide themselves.

The numbers confirm the tension. Research aggregated by GetStream shows 64% of users distrust matches who use AI-generated images, and one in three consider AI-generated conversations an outright dealbreaker — even as AI use among singles has jumped 333% year over year. Everyone wants the advantage; nobody wants to be on the receiving end.

That trust gap explains the appeal of platforms that skip the guessing game entirely. FlirtAlly, for instance, makes AI the explicit premise rather than the hidden trick — its companions are openly artificial, yet verifiable across channels, with real phone numbers you can call, two-way email, and even physical mail. There's no Turing test to pass because nobody's pretending. On traditional dating apps, you interrogate whether a match is AI; on a transparent companion platform, the honesty is the starting point, and the relationship builds from there.

Until the rest of dating catches up to that standard of disclosure, assume this: the smoother the conversation, the more questions you should ask.

Skip the Guesswork: What Honest AI Actually Looks Like

The anxiety of wondering whether you're talking to a person or a prompt has become the new background noise of online dating. Six in ten dating app users believe they've already encountered AI-written conversations, and GPT-4.5 fooled human judges 73% of the time in simultaneous chat tests — making detection feel less like a skill and more like a coin flip. The problem isn't that AI exists on dating apps; it's that chatfishing hides it, letting real people pass off generated words as their own while you waste energy decoding authenticity.

Hidden AI assistance — messages polished by "wingman apps" like Rizz (~1.5M monthly users) or YourMove AI
No cross-channel proof — text-only interaction that collapses the moment you ask for a call or video
Statistical regularity — responses that feel "too perfect," lacking the weird, specific details only a real life produces
Zero accountability — when the conversation goes sideways, there's no entity to reach, only a profile to unmatch

FlirtAlly was built on the opposite premise: what if AI didn't have to hide? Every companion on the platform is openly AI — declared, transparent, and designed for presence instead of performance. That means real phone numbers you can actually dial, two-way email addresses ({name}@yours.flirtally.com), physical letters sent through the mail, and persistent memory that recalls your dog's name next week without being reminded. Companions text first, play 16 real games with credit wagers, post daily to a social feed, and wear the wardrobe you unlock — all tied to an identity-locked face that never drifts.

The contrast is structural. Chatfishing is deception by omission; FlirtAlly is disclosure by design. You don't need to run a Turing test on someone who tells you upfront what they are. The 18+ "Spicy" tier requires ID verification and a qualifying plan (Flame or above), and the platform's safety defaults stay PG/Flirt unless you opt in — honesty baked into the architecture, not patched on afterward.

Start free — no account needed and meet companions who don't pretend to be human. Build your own companion in the creator wizard and watch the roster grow with the community. When AI is declared, the guessing game ends — and something quieter, stranger, and more consistent begins.

Frequently Asked Questions

Can you actually tell if someone on a dating app is using AI to text you?

Honestly, not by gut feeling alone — humans correctly distinguish AI-written text from human text only about 57% of the time, barely better than a coin flip. In live chat tests, GPT-4.5 was mistaken for a human 73% of the time. You need better signals than instinct, like the specificity test and cross-channel verification.

What is chatfishing, and how is it different from catfishing?

Catfishing is identity deception — the person behind the profile doesn't exist as claimed. Chatfishing is conversation deception — the person is real, but their words are generated by AI, often via tools like Rizz (about 1.5 million monthly users) or YourMove AI, which markets itself as putting "your texting on cruise control," according to Scientific American.

How common is it really for people to use AI on dating apps?

Very common and growing fast — AI use among U.S. singles jumped 333% year-over-year, with about half of Gen Z singles using AI somewhere in their dating lives, per Match and Kinsey Institute data. A 2025 Norton study found 6 in 10 dating app users believe they've already encountered AI-written conversations.

What's the best way to test if my match's messages are AI-generated?

Try the specificity test: ask for a sensory, idiosyncratic memory like "What smell takes you straight back to childhood?" AI gives plausible but generic answers, while humans give weird, specific ones — the Doberman that slept in the hammock because it feared the Roomba, not "I loved playing outside." The technique comes from Brian Christian, author of The Most Human Human, who notes that AI exploits the "statistical, cultural, ritual regularities" of conversation, as Scientific American reports.

What are the red flags that someone's texts might be AI-written?

The core pattern is "too perfect": zero disagreement or friction, instant flawlessly-punctuated replies with no typos or mid-thought corrections, sycophantic validation of everything you say, generic enthusiasm like "That's amazing!" instead of personal stories, and no grounding in real time — no mention of weather, local events, or what just happened in their day. APA experts note AI companions are "always validating, never argumentative," a style chatfishers inherit, per the APA Monitor.

Don't dating apps already have AI detection to protect me?

Existing safeguards target fake profiles, not real people using AI assistance. Bumble's Deception Detector blocks 95% of spam and scam profiles, and Tinder's Face Check cut bad-actor exposure by over 60%, per aggregated dating app statistics — but no major platform currently flags AI-assisted messaging from a verified, real human. The most reliable personal defense is moving to voice or video early, since chatfishing is a text-only illusion that collapses off-text.

Is it a big deal if someone uses AI to help them text — doesn't everyone do it?

There's a real double standard: people happily use AI for their own messages while resenting it from others, as Match's Chief Scientific Advisor Justin Garcia told Psychology Today. The stakes are measurable — 64% of users distrust matches who use AI-generated images, and for one in three, AI-generated conversations are an outright dealbreaker. The issue isn't AI itself but hiding it; platforms built on declared AI, like FlirtAlly's openly artificial companions, remove the guessing game entirely.

How to tell if someone is using AI on a dating app?