OpenAI CEO Sam Altman recently said that the new update of ChatGPT feels "like magic," and this evaluation is spot on. With this update, OpenAI has sent a clear challenge signal to the tech giants: a new round of AI competition has begun.

On Monday, OpenAI CTO Mira Murati showcased the "Spring Update" of ChatGPT through a series of live demonstrations. Powered by the GPT-4o model, this AI chatbot can now perform real-time reasoning between audio, visual, and text, delivering impressive performance.

The new version of ChatGPT has made significant breakthroughs in sound and conversational capabilities. It not only sounds emotional but can also change its tone effortlessly. In the demonstration, ChatGPT's voice resembled that of an American woman, reminiscent of Scarlett Johansson's voice in "Her," although OpenAI researchers also made it switch to a robotic voice at some point. According to an OpenAI spokesperson, audio output will be limited to a range of preset sounds upon release.

This AI not only has realistic voice but also demonstrates astonishing ability to imitate human speech. The new ChatGPT can laugh, add humor based on prompts, and flexibly adjust the intonation of its voice. It can even pick up subtle cues from humans. In one demonstration, when a researcher took a deep breath, ChatGPT cleverly responded, "Mark, you're not a vacuum cleaner."

In addition, ChatGPT allows users to interrupt its speech, making the conversation more natural and fluent. Users can clarify questions or change topics without waiting for the AI to finish its response.

Impressively, ChatGPT also has exceptionally fast response times. According to an OpenAI spokesperson, this chatbot can process audio input at a response rate close to that of humans, with an average response time of only 320 milliseconds.

With this major update of ChatGPT, OpenAI once again showcases its leading strength in the field of AI and brings new challenges and opportunities to the tech industry.