⚡Megatron LLM Hub
←Back to search
Twitter

手机 Ad Guard 与 SecurityRisk 组件对非通讯录号码及微信/QQ 通话触发 ASR 转写并外传。

给出预装组件触发条件的逆向细节,可直接用于移动端取证与合规评估。
移动端隐私逆向
But when does the system decide to monitor a call for 'fraud' you may ask !? Well, according to the code in the phones 'Ad Guard' and SecurityRisk components, it latches on to calls that are not in the user's phonebook or not matching the 'yellow pages' database. Then a ASR transcriber sets in and it transmits the call's content for classification. If the call is VoIP from WeChat/QQ it always gets transmitted it seems ( either video or just audio, depending on what is active ). It has to be said, that this functionality can be disabled by the user. There is a debug variant of it built in, that doesn't read the 'has agreed to have the calls monitored' flag. But overall it needs the consent of the user to actually work. As it was stated in the earlier post, the system also transmits a lot of other details on the user and the connection to Tencent by default. And if the story is true, it seems to be able to get directly accessed by the Police's automatic analysis systems. Which results in calls from the law enforcement if illegal app's are used on a phone. Technically this feature is emulating exactly what the official anti-fraud app of the Ministry of Public Security released a while back.