You can earn money by speaking dialects? WeChat (微信) invites some users to participate in voice data collection; red packet rewards credited directly

Quick take

WeChat (微信) has reportedly begun inviting a subset of users to a “dialect collection” activity that pays small cash rewards for voice recordings. The invitation arrives as an official push from the “WeChat Team (微信团队)” and directs invitees into a mini‑program where they read short, everyday sentences in local dialects. Rewards are said to be delivered as red packets credited directly to users’ WeChat Wallet (微信零钱) after review.

How the program works

According to user reports collected by IT之家 (IT Home), participants read prompted sentences in their dialect and submit recordings; roughly three sentences earn about ¥1, 20 sentences about ¥5, and some users say the system allows 100–200 sentences a day — a theoretical cap that could yield up to ~¥40 in a day. The platform reportedly requires clear, genuine recordings and runs anti‑cheating checks; payments are paid within 30 days after audit. It has also been reported that some users saw the activity already closed after a short run.

Why it matters

Why would Tencent’s WeChat collect dialect speech? Improvements to voice‑to‑text and speech‑recognition models are the likely goal, and the feature could also feed efforts to document and preserve China’s many regional languages. WeChat’s voice‑to‑text already supports certain dialects, reportedly including Teochew (潮汕话). The push comes amid broader concerns about dialect loss: surveys cited by Chinese media show many dialects have dwindling speaker bases, and a National People’s Congress deputy, Luo Tian, reportedly urged stronger preservation measures at the 2026 “Two Sessions.”

Context and considerations

For Western readers: this is part of a wider trend in which large Chinese tech platforms collect domestic voice and text data to train AI models while export controls and sanctions have limited access to some foreign chip and model supplies. That raises two linked questions — linguistic preservation and data governance. User reports are the primary source for this short pilot, so many details remain unconfirmed. Reportedly, rewards and the invitation mechanism were limited and selective, and privacy and usage terms for submitted voice data should be reviewed by any participant before joining.