A few days ago Claude AI releases new Computer Use featureLetting AI autonomously control computers – recently, China’s Zhipu AI also launched a similar feature. Zhipu AI released GLM-4-Voice, an end-to-end emotional voice model. GLM-4-Voice supports multiple languages, low latency, and can be interrupted at any time. It’s now available in Zhipu Qingyan, Zhipu AI’s artificial intelligence application. In addition, Zhipu AI also launched a tool called AutoGLM, which can accomplish anything a human can do on a phone. AutoGLM’s phone use feature can receive simple text/voice commands from users and mimic how humans use their phones. It’s fair to say that the phone use feature from Zhipu AI works on similar principles to Claude AI’s Computer Use – except Claude AI operates the user’s computer, while Zhipu AI operates the user’s phone. Today we’ll introduce what Zhipu AI’s phone use feature can actually do. If you’re interested, keep reading~
No need to use your hands! Chinese AI companies have launched a service that lets you easily control your phone with your voice.
Chinese AI company Zhipu AI announced on 10/25 that its large model family has a new member: GLM-4-Voice, an end-to-end emotional voice model. GLM-4-Voice can understand emotions, express feelings, and create emotional resonance. It supports multiple languages with low latency and can be interrupted at any time. Zhipu AI’s GLM-4-Voice should perform similarly to ChatGPT’s advanced voice mode. In addition to launching the new large model, Zhipu AI also introduced something new on the tool-usage front: AutoGLM. Through GLM-4-Voice and AutoGLM, Zhipu AI is driving progress in multimodal interaction and tool usage. So what exactly does AutoGLM do? Keep reading to find out!
Zhipu AI hopes that in addition to launching large models, it can also improve the tool capabilities of its models, so they thought of the device users interact with the most: the phone. AutoGLM has a “phone use” feature. According to the official demo, you only need to give AutoGLM simple text or voice commands, and AutoGLM can simulate human operations on the phone. Based on the command, it opens the corresponding app and performs the required actions. For example: ask AutoGLM to like your boss’s Moments post on WeChat and leave a comment, give a five-star review on a restaurant rating app, shop on Taobao, and more. You can watch the video below to learn more:
AutoGLM’s mobile version is now in closed beta, and registered users can already experience AutoGLM’s phone use feature. Some netizens have already shared videos of themselves using AutoGLM’s phone use function to find and book the nearest, most cost-effective hotel near a scenic spot.
Send message using AutoGLM:
Use AutoGLM to send a video from the photo album to a WeChat friend:
If you want to order delivery, that works too—just tell AutoGLM you want an Americano, and it will automatically open the delivery app and find the store for you. Not only that, AutoGLM will also ask you which of the two nearby stores you’d prefer, and whether you want a regular Americano or another flavor. You could say it truly achieves a hands-free experience.
AutoGLM’s phone use feature is currently in beta testing and temporarily only supports Android systems. No information has been disclosed yet regarding when it will officially launch.
Zhipu AI has launched two major updates this time: GLM-4-Voice and AutoGLM. GLM-4-Voice focuses on voice interaction, delivering emotionally rich voice conversation experiences. AutoGLM, on the other hand, focuses on AI’s ability to use mobile phone operation tools. Both are indispensable elements for Zhipu AI’s future transition to Artificial General Intelligence (AGI). For those interested in learning more about Zhipu AI’s update information, you canGo to the Zhipu AI official website。
Source: KOCPC Chinese