OpenAI's ChatGPT gets some massive upgrades, now lets people talk to ChatGPT

Sept. 30, 2024, 6:35 a.m.

Read time estimation: 4 minutes.

  2

ChatGPT is stepping up its game with a brand-new voice feature, making conversations with the AI more natural and fluid. OpenAI announced that its popular chatbot can now engage in audio chats, but for now, it’s only available for those who subscribe to the premium service.

This advanced voice capability provides a more seamless experience, allowing users to have real-time discussions. The AI even pauses if interrupted, adding a touch of conversational realism to the chat.

While this new voice feature is being gradually rolled out this week, it is not yet available in the EU, the UK, and some other European countries. OpenAI initially announced plans for the voice feature back in May, which generated significant buzz due to a voice remarkably similar to Scarlett Johansson's from the film Her.

However, legal action quickly halted the use of that voice. Since then, users on the free version have been able to experiment with other voices. The premium version offers a wider selection of nine voices and allows users to customize their instructions for these chats within the app's settings.

In a playful nod to the wait, OpenAI’s co-founder and CEO Sam Altman posted on X (formerly Twitter), saying, “Hope you think it was worth the wait.”

Rising Competition in AI Voice Tech

OpenAI is not alone in the AI voice game. Google recently launched their Gemini Live voice feature, and Meta is also entering the fray with plans to release celebrity voices on platforms like Facebook, Instagram, and WhatsApp later this week.

This competitive environment emphasizes the growing significance of AI-powered voice capabilities, particularly for tech giants. OpenAI, backed by Microsoft, continues to lead the way, having established a substantial advantage with ChatGPT since its introduction in late 2022. The chatbot has already amassed over 200 million weekly active users, according to an August report.

However, access to the voice feature is restricted to subscribers of OpenAI's Plus, Team, or Enterprise plans, with the most affordable option priced at $20 per month.

Upgrades for GPT-4o Mini

In addition to voice capabilities, OpenAI has introduced several noteworthy updates for its smaller GPT-4o mini model. Previously considered less potent, the mini model has now been enhanced to offer four key features previously exclusive to the larger GPT-4o version.

Firstly, the mini model can now generate images from text prompts using the DALL-E 3 model, mirroring its larger counterpart. This upgrade is anticipated to be well-received by users seeking faster image generation without compromising quality.

In another significant upgrade, the smaller model now has the ability to access the internet, allowing users to retrieve up-to-date information and conduct real-time research. This enhancement brings it closer to the capabilities of the larger GPT-4o, providing users with greater flexibility for fact-checking and information gathering.

Another new feature is the ability to upload and analyze documents and images, which will make working with complex visual data much easier. Users can now interact with both text and visuals, paving the way for educational and personal uses.

Finally, the GPT-4o mini model can now remember previous conversations with users, similar to its more advanced counterparts. This memory feature allows the model to provide more relevant follow-up responses and recognize user preferences, making long-term interactions more seamless.

Advertisement