• Announcements

  • Android 15 QPR2 / Android 16 Beta microphone indicator persistence bug

Android 15 QPR2 introduced a bug where the Microphone indicator will sometimes remain active after Microphone usage ends. We've confirmed this issue is present in the stock Pixel OS for both Android 15 QPR2 and Android 16 Beta 3. We're working resolving the regression but haven't figured it out yet.

Here are several upstream issue reports:

https://issuetracker.google.com/issues/388151378
https://issuetracker.google.com/issues/392596949
https://issuetracker.google.com/issues/401832184

It does not mean that apps are actually continuing to use the Microphone. They introduced a bug where the OS can miss that it stopped.

I see that the first issue you linked (#388151378) has now been marked as fixed.

    @n3t_admin "if your doctor is using WhatsApp, he might be violating data privacy laws by doing so. Depending on your country. I would consider looking for a different doctor; Meta has no business knowing my medical history."

    Are you implying that meta can directly read whatsapp messages? I willfully engage in conversation with everyone I message on whatsapp, and I take full responsibility for what might result from that, just as I would from any other chat app.

    My biggest complaints about whatsapp are the memory vulnerabilities, that make it much easier to exploit than other messaging apps. It's not fully open source, the metadata it collects, etc... But I try to limit the potential damage it might be able to do, that's why I run it on GrapheneOS.

      5 days later

      locked The new "KI assistant" can read all (WhatsApp) messages on the phone. All chats done with the KI assistant will apparently be stored on the Meta servers. The KI assistant will also automatically activate the microphone when opened and will listen to everything.

          abdullah

          It inferred that from this from this article

          Additionally, when users attempt to search for a chat, group, or contact, they will encounter AI-generated suggestions and prompts that integrate Meta AI responses.

          If the assistant can give suggestions based on chats, then I would assume it reads the contents of those chats. But maybe it only looks at the meta data.

            Would this bug cause a microphone to be green in use after a reboot?

            How do we not know that people reporting this error are not just covering for an actual exploit being done?

            I've seen the green dot now after a reboot when not using any Apps. I don't trust this is just some random error.

              angela Would this bug cause a microphone to be green in use after a reboot?

              No.

              angela How do we not know that people reporting this error are not just covering for an actual exploit being done?

              We know this particular bug talked about in this thread is not an exploit, since GrapheneOS developers themselves has been able to reproduce the bug. That doesn't mean there isn't also exploits going on. But in the cases where exploits are going on, it is good the users get alarmed by microphone dot showing up. That means that security system is working as it should.

              But we rather not people get alarmed when there is no reason to, such as by a harmless bug such as in the case talked about in this specific thread.

              angela I've seen the green dot now after a reboot when not using any Apps. I don't trust this is just some random error.

              You mean before you enter your passphrase or PIN for the first time after the reboot? You mean in before first unlock (BFU)? Theoretically, an app can run in BFU if it registers itself in a very specific way. Maybe you have a malicious or compromised app installed? An app cannot use the microphone without the dot showing up after all. But if your system was entirely compromised, I would have expected the attacker to make sure the dot doesn't appear even if they use the microphone.