OpenAI has finally addressed a major problem with ChatGPT’s voice mode feature

ChatGPT has made an important change to its interface, making voice mode more easily accessible than before.

for its smartphone app and web interface. Then, in March last year, the company introduced real-time conversation capabilities to it. When you tap the waveform icon next to the ChatGPT text field, you’re taken to an orb-filled interface where you can chat with the chatbot in real time. That natural, real-time conversation ability still exists, but now, instead of taking you to a separate orb-filled interface, ChatGPT lets you use voice mode directly within the chat.

Tap the waveform icon to use your voice command, and you would be able to communicate with ChatGPT directly in the chat. This change has been made available in both the ChatGPT app and the web interface. You might wonder, what was even the need for such a small change?

In the orb interface, you could only listen to what ChatGPT says instead of seeing its transcript on the screen. So if you didn’t clearly hear what ChatGPT said, you had to leave voice mode to check the transcript and find out the response. But now, since the voice capability has been moved to the chat interface, you can easily view the transcript while talking with ChatGPT.

OpenAI’s chatbot can now also display real-time information such as maps, weather updates, and more. Unfortunately, the feature didn’t perform as I hoped it would. Although the chatbot accurately described the weather in my area, the map functionality didn’t work as I expected.

When I asked for a map showing the top restaurants in my area, I was given links to the directions of those restaurants instead of seeing the map interface within the chat window, as OpenAI demonstrated in the video posted on X. I even narrowed down my request by asking to show the map of a specific restaurant, but I was still provided with a link to its directions instead of the map interface in the chat window. Strangely, the map interface did appear when I used the exact same command shown in the video posted by OpenAI.

More people would likely use the voice mode

ChatGPT logo on a smartphone. | Image by PhoneArena

When ChatGPT’s voice mode was launched, it felt like a very interesting feature, and users were having a great time using it. But over time, people stopped using it as much as they did when it first came out. I also can’t remember the last time I used it to search for anything.

But now, since the feature has become more accessible after being moved directly into chat, OpenAI would hope more users will start using it. It’s also worth noting that the more users interact with this feature, the more voice commands OpenAI will have to train its models.

OpenAI will likely find itself in a comfortable and advantageous position after moving voice capabilities to the chat window. Fortunately, you have the option to stop OpenAI from training its AI models on your voice. Here’s how you can do that:

Open the ChatGPT app on your smartphone, tap the Customize icon in the top-left corner, and then tap your name to open the ChatGPT settings window
Select Data Controls
Turn off the toggle next to Include your audio recordings

That said, not everyone may prefer the voice mode in the chat interface. That’s why you also have the option to switch back to the previous orb mode interface. To do that, open the ChatGPT settings window on your smartphone, select Voice, and enable the Separate Mode toggle. On the ChatGPT web interface, open the ChatGPT settings, select Personalization, click Advanced, and then turn on the Separate Mode toggle.

A Better ChatGPT

By integrating the voice mode into the chat interface, OpenAI seems to be aiming to make interactions with the chatbot feel more natural. You can now easily switch between text and audio modes within the same conversation in ChatGPT. However, I noticed that the voice mode feature continues to listen and respond to your speech until you manually hit the End button.

As a result, it once happened to me that I used the voice mode feature and forgot to press the End button. About five minutes later, I asked my mother to make tea for me, and ChatGPT thought I was still talking to it and started telling me the recipe for making tea.

I really hope OpenAI introduces a feature that automatically deactivates voice mode after a certain period of inactivity. That said, ChatGPT’s rival, Gemini, already offers a similar feature called Gemini Live, which switches to a completely new screen when you use it. Gemini, however, has long offered a transcript button in the top-right corner of the Gemini Live window, letting you view the transcript while also hearing what the AI has to say.

I used to rely heavily on Gemini Live on my Pixel 10 because it could answer me through speech while displaying the transcript in real time. Now that this feature is available in ChatGPT as well, I might switch to it since I prefer ChatGPT’s voice experience over Gemini Live.

Trending Products