OpenAI has just rolled out its Advanced Voice Mode in ChatGPT across the European Union, bringing it to users in Switzerland, Norway, Iceland, and a few other regions. The feature was previously limited to the U.S. and U.K., but now more ChatGPT Plus and Team subscribers can access it. OpenAI continues to push its voice interaction capabilities as it works to expand global AI services. With this feature, conversations with AI get a real-time, voice-driven update.
Well yes.
All Plus users in the EU, Switzerland, Iceland, Norway, and Liechtenstein now have access to Advanced Voice. https://t.co/AC6mdd1Lkf
— OpenAI (@OpenAI) October 22, 2024
AI Voice Gets Faster and More Personal
Voice Mode is quick, really quick, delivering audio responses in under a third of a second. That’s about 232 milliseconds, according to some sources. The new feature was initially tied to the GPT-4o launch but has since evolved.
However, there’s a daily usage cap. If you’re getting close to the limit, the app will warn you, letting you know when you’ve got just 15 minutes of voice use left. The limit keeps things balanced across the board for all users, though not everyone loves having a cap.
Competition in AI Voice Heats Up
This launch arrives as Apple and Google’s AI offerings struggle to keep pace. Siri, for example, is underperforming—by a lot. Leaked internal documents from Apple, as reported yesterday, show that Siri lags far behind ChatGPT in key areas, including accuracy. Siri’s performance has been 25% lower compared to ChatGPT’s, and OpenAI’s model answered more queries correctly.
Apple faces a tough challenge. While Siri can run on powerful hardware like the new iPad mini, the software needed to fully utilize that hardware is still catching up. Meanwhile, Google’s Gemini Live—launched in September—has its own problems. Though it allows users to speak to the AI in real-time, users report it lacks some basic personalizations, making it feel a bit robotic at times.
Customization and Voices in ChatGPT
OpenAI has been adding new features to ChatGPT’s Voice Mode, which makes it stand out. Five new voices—Arbor, Maple, Sol, Spruce, and Vale—are available now, and there’s more focus on accents and smoothness in foreign languages. The total number of available voices is now up to nine. Users can also customize their ChatGPT experience by giving the AI specific instructions to follow in conversations, allowing for a more tailored experience that feels more personal than its competitors.
In comparison, Google’s Gemini Live isn’t quite there yet. It can’t adjust tone or pitch, and users have been frustrated with its lack of natural flow during conversations. Google has promised more updates, but for now, OpenAI appears to have the upper hand when it comes to personalization.
Daily Usage Limits and Access
For those wanting to jump in, all you need is the latest version of the ChatGPT app from either the Play Store or App Store. But be aware—there’s a daily limit on how much voice interaction you can have with the AI. Once you’re nearing the limit, the app gives you a heads-up, notifying you that you’ve got 15 minutes left of chat time. While some users might find this a bit restrictive, it helps ensure smooth service for everyone.
The voice feature was first made available in the U.S. and U.K., and this rollout expands it further into Europe, making it available to even more users. As the voice interaction space grows, OpenAI continues to push for global accessibility.
Voice Mode’s Tech: Quick and Reliable
OpenAI’s goal with Voice Mode was to make AI conversations as smooth and natural as possible. By cutting down response times to under a second, the company has managed to create a system that feels much more human. The voice responses clock in at about 320 milliseconds on average, making the interactions nearly as fast as a regular conversation. The feature has been designed to handle various accents better, which will likely be a hit in Europe, where linguistic diversity is key.
In contrast, Apple’s Siri and Google’s Gemini Live both struggle with these fast, real-time interactions, making OpenAI’s offering stand out. Siri, in particular, has been in the news for its disappointing performance, especially after documents leaked showing how far behind it is compared to OpenAI’s models.
The Bigger AI Race
Apple and Google are both trying to catch up. According to a report published just yesterday, Siri is currently answering 30% fewer questions successfully than ChatGPT. Apple’s own employees are worried about this AI gap, highlighting concerns over how far behind they’ve fallen in this race.
Meanwhile, Google’s Gemini Live also faces challenges. Despite being available for a few months now, it’s still struggling with delivering accurate, personalized responses. Google had initially hoped to position Gemini Live as a direct competitor to ChatGPT, but user feedback suggests it’s falling short.
OpenAI’s Reasoning Models
Beyond voice, OpenAI has also been focusing on improving reasoning capabilities. The company introduced its o1 model in September, designed to better handle tasks requiring complex problem-solving, such as coding or multi-step logical challenges. This model builds on the “Strawberry” AI methodology and incorporates reinforcement learning, a different approach from OpenAI’s earlier models.
This reasoning power helps explain why ChatGPT is leading the pack in AI voice interactions. Though Google and Apple are working on their own improvements, it seems OpenAI’s unique training methods give it an edge for now.
AI Safety Delays and Concerns
It’s not all smooth sailing for OpenAI, though. Back in June, the company had to delay the launch of a voice assistant feature due to safety concerns. OpenAI has also faced criticism over the potential misuse of voice cloning technologies, which can replicate human speech using just a 15-second sample. After facing backlash, the company has proceeded cautiously with voice feature rollouts, conducting more safety and performance checks before launching it more widely.
At the same time, the Gemini Live feature from Google, introduced in September, was also hit with performance issues, making this AI space an ongoing battleground for tech giants. Both Google and Apple are focused on addressing concerns about hallucinations—where AI confidently presents incorrect information.
While these companies work on refining their AI, OpenAI is leading the charge with its fast, customizable voice interactions, making ChatGPT’s Advanced Voice Mode the tool to beat in the AI conversation space.
Last Updated on November 7, 2024 2:24 pm CET