Cant turn off setting for ChatGPT listening while it talks in voice mode after r
On April 13th, someone in the ChatGPT community forum noticed something missing. Not a crash, not a degraded response — a vanished control. In voice mode, there had previously been a setting that let users decide whether the microphone stays active while ChatGPT is speaking. The user reported that after a recent update, that toggle was gone. The setting, as best as the post describes it, governed whether the app continues listening during its own output — a distinction that sounds technical until you sit with it for a moment. An app that listens only when you speak is a voice assistant. An app that listens while it talks is something closer to an open microphone. The original thread hasn't generated a documented fix, an official response from OpenAI, or a chorus of corroborating voices. One signal, holding its position since mid-April.
If confirmed, here is what this means. OpenAI quietly removed user agency over a microphone behaviour that carries real privacy weight. The people most affected are anyone using voice mode in shared or sensitive environments — offices, homes with other conversations happening nearby, anywhere ambient audio has value or vulnerability. The second-order effect is subtler: when companies remove granular privacy controls without announcement, they don't just change a feature, they shift the terms of trust. Users who built a mental model of the app's listening behaviour around that toggle now have no reliable baseline. For OpenAI specifically, this arrives at a moment when the industry is already absorbing sustained scrutiny over how AI products handle audio data. A silent removal of a microphone control isn't a neutral product decision — it reads, fairly or not, as a choice about what users are allowed to know.
Watch for other users in the ChatGPT forum confirming the same behaviour after the same update window, or any official changelog from OpenAI that either restores the setting or explains its removal.
<!-- SC_OFF --><div class="md"><p>Haven't seen anyone else report this, so maybe I'm a part of some A-B testing, but about 12 hours ago, I noticed that standard voice mode can be
<!-- SC_OFF --><div class="md"><p>I've been using chatgpt to produce files for me where I give an example variable and have it iterate the example over hundreds of variables. When
<!-- SC_OFF --><div class="md"><p>For me, when I select GPT-5.5 Thinking with Auto-switch off; standard voice generates a response labelled GPT-5.3. If I leave voice mode and use
<!-- SC_OFF --><div class="md"><p>As the title says. I'm getting absolutely no thinking time reported on my prompts, and the answers are coming out completely instant. It feels
<!-- SC_OFF --><div class="md"><p>​</p> <p>I have a quick question regarding my recent experience with ChatGPT.</p> <p>I recently renewed my subscription after a two-month
<!-- SC_OFF --><div class="md"><p>The visceral reaction that my brain has to hearing a robotic voice pretending to breathe and make mouth noises is horrifying. The Uncanny Valley
<!-- SC_OFF --><div class="md"><p>Earlier today and all other times I've used it, while I had it set to Pro, and it's currently thinking, I could give it input using the text box.
<!-- SC_OFF --><div class="md"><p>lists, numbering, tables... it can't articulate 2 full sentences.. like i can't read, and i need an analogy for everything, it feels like text
NewsHive monitors these sources continuously. All signal titles above link to the original reporting.
Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →