Call-recording app Neon, which became the #2 social app on the US App Store, goes dark after exposing users' phone numbers, call recordings, and transcripts
A viral app called Neon, which offers to record your phone calls and pay you for the audio so it can sell that data to AI companies …
TechCrunch
Context & Ripple Effects
Neon’s shutdown follows its rapid ascent after it offered users payment in exchange for call audio intended for AI training, a model documented in coverage of Neon’s data-for-audio exchange. The exposure turns that growth story into a test of whether a consumer app can make highly sensitive recordings a viable data supply source.
The episode also sits alongside audio products built around recording and transcription, including Airchat’s recorded-and-transcribed social feed. What distinguishes Neon is that call content, phone numbers, and transcripts were reportedly exposed while the app’s business model depended on collecting that material.
First-order effects
Neon is unavailable, interrupting its ability to collect call recordings and offer users compensation for them.
People whose phone numbers, recordings, or transcripts were exposed face an immediate privacy risk; the exposure also undermines confidence in Neon’s handling of the data it sought to sell.
Second-order effects
Apps that collect, transcribe, or monetize voice data will face closer scrutiny of how recordings and associated contact data are secured, especially when their value proposition asks users to trade sensitive data for payment.
Potential buyers of consumer-sourced voice data may have to weigh the reliability and handling of the underlying collection channel more carefully after Neon’s shutdown.
Third-order effects
If consumer AI-data collection continues to move toward direct compensation for personal content, security and disclosure practices may become a core constraint on whether such marketplaces can operate at scale.
The case sharpens the public-data permission boundary: consent to record or sell one participant’s call data may not resolve the privacy interests of everyone captured in the conversation.
The trend: Neon is a data point in the push to turn consumer-generated voice data into AI-training supply, with privacy and data stewardship becoming decisive limits on that model.
Humanity's last use case is to generate training data. “the No. 2 social app on the Apple App Store, pays users to record their phone calls and sells data to AI firms” https://techcrunch.com/...
Neon Mobile, now the No. 2 free social-networking app on the U.S. App Store, pays users $0.30 per recorded call minute (capped at $30/day) and then sells the audio to AI firms. Neon claims it only captures the caller's side unless both parties use the app, but its terms grant a …
ICYMI: A viral app called Neon, which pays you to record your phone calls so your audio can be used to train AI, has gone offline after I discovered a security lapse that allowed any user to access the phone numbers, call recordings, and call transcriptions of any other user.
This, and the similar Tea data leak, is a reminder that just because it's in an official app store, doesn't mean an app is good or secure. — techcrunch.com/2025/09/25/v...
The solution to the male lonliness epidemic is getting men to call each other on the phone so they can sell their voices/speech for AI training techcrunch.com/2025/09/24/n...
ICYMI: A viral app called Neon, which pays you to record your phone calls so your audio can be used to train AI, has gone offline after I discovered a security lapse exposing users' phone numbers, call recordings, and call transcriptions. — Anyone could download users' raw audi…
“Neon, the No. 2 social app on the Apple App Store, pays users to record their phone calls and sells data to AI firms” — 30 cents a minute ($18/hr) when you speak with other Neon users. Referral fees as well. — Feels like we lost the thread, somehow. But better than using p…
There's no such thing as anonymous personal data. Especially not when it's the story of your life as recorded in your phone conversations. — AIs can make voices that sound exactly like you with a tiny sample of your recorded voice. That capability can make it possible for cri…