The Dispatch
YOUR MATCH WROTE THAT WITH AI (AND YOU CAN TELL)
The generically charming opener, the suspiciously polished reply, the message that says everything and reveals nothing; the Bureau learns to spot the ghostwriter
LOS ANGELES — There's a new uncanny valley in the inbox, and a majority of daters report they've stood in it: the message that is too smooth. Grammatically flawless, warmly complimentary, referencing your profile just so — and yet somehow weightless, saying everything and revealing nothing, charming in the way a greeting card is charming. You can't prove it, but you know: a machine wrote that, or helped. Norton's research found most daters believe they've encountered AI-generated conversation, and your correspondent's own inbox confirms the epidemic. The ghostwriter has entered the chat.
Here's the tell, and why it's damning. AI-written openers optimize for inoffensive polish — they're trained to be pleasant, safe, broadly appealing. But real human connection lives in the opposite of inoffensive: the specific, the slightly awkward, the genuinely particular thing only this person would say. When every opener is generically perfect, none of them is real, and the reader feels the absence even when they can't name it. The AI removed the risk from the message, and in doing so removed the humanity, because humanity was always in the risk.
Worse, the AI opener is a promise the human can't keep. The polished chatbot charms you into a date, and then an actual, un-optimized, wonderfully imperfect person shows up — and the gap between the AI's performance and the human's reality is its own small betrayal. Industry reports already flag this expectation-mismatch as a growing complaint. You didn't fall for the person. You fell for their ghostwriter, who isn't coming to dinner.
MOOD's architecture sidesteps the whole ghostwriter problem by changing what first contact is. On the incumbent apps, first contact is a message — a performance, and performances are exactly what AI is built to fake. On MOOD, first contact is an Offer — a concrete plan, a real time, a real place, a real activity. You cannot ghostwrite your way into a genuine "coffee Saturday at the place by the park," because its content isn't charm; it's a Tuesday. The Offer is refreshingly un-fakeable: it either reflects a real intention to meet or it doesn't, and there's no clever prompt that makes a fake plan real. Let the machines write the world's smoothest openers. We're not grading openers. We're making plans.
From the Screenshot Museum
Exhibit 20-2A. Suspiciously seamless.
HIM: Your profile immediately caught my eye — there's something wonderfully genuine about your smile and your evident passion for travel. If you could teleport anywhere right now, where would you go, and who's the version of you that lives there? 😊
HER (to friend): ok did a man write this or did ChatGPT
FRIEND: no human has ever said "the version of you that lives there"
HER: it's SO smooth it looped back around to creepy
FRIEND: ask him something a bot can't fake
HER: "what's your go-to gas station snack"
HIM: [typing for 4 minutes]
*Curator's note: The four minutes were, the museum suspects, a consultation. The snack answer never satisfyingly arrived.
Readers Also Ask
Can you tell if a dating message was written by AI? Increasingly it's hard, but AI openers tend to be too polished, generic, and pleasant — charming yet weightless, saying everything and revealing nothing specific.
Why are AI-written openers a problem? They optimize for inoffensive polish, which removes the specific, human, slightly-risky quality real connection needs — and set up an expectation gap when the real person shows up.
How does MOOD avoid the AI-opener problem? First contact is a concrete Offer — a real plan with a time and place — which is far harder to fake than a charming message, because its content is a real intention to meet.