Manus: China’s New AI Tool Sparks Excitement

A powerful new AI agent, Manus, is gaining attention in China, raising hopes that it could follow the success of DeepSeek—the advanced chatbot that shook the global tech industry earlier this year. Unlike traditional chatbots, Manus is designed to perform complex tasks autonomously, from stock market analysis to trip planning, according to its developers.

Rapid Rise of Manus

The new AI programme was launched last week by the Chinese startup Butterfly Effect. In a viral introduction video, co-founder Yichao “Peak” Ji described it as “the next paradigm of human-machine collaboration” and even hinted at its potential to be a step toward Artificial General Intelligence (AGI)—AI that thinks like humans.

Despite being accessible only by invitation, Manus has quickly gained popularity. Its Discord community has already grown to over 170,000 members.

The name “Manus” is derived from the Latin phrase “Mens et Manus”, meaning “mind and hand”, reflecting its ability to combine knowledge with action.

Exclusive Access and Unique Features

Unlike DeepSeek or ChatGPT, which primarily generate responses in a chat-based format, this functions as an autonomous AI system. It can execute tasks such as booking tickets, sorting resumes, and even creating custom websites.

AI researcher Manoj Harjani from Singapore’s S. Rajaratnam School of International Studies (RSIS) highlighted this distinction, noting that Manus is designed to act rather than just respond.

However, the tool’s invite-only beta phase could impact its long-term success. Marina Zhang, an associate professor at the University of Technology Sydney, warned that limiting access might slow down adoption. In contrast, DeepSeek’s open-source engagement helped it build a massive user base.

When AFP tested Manus, the AI took longer than DeepSeek to generate responses but was capable of handling more complex tasks.

Handling of Sensitive Topics

One major difference between Manus and other Chinese AI models is its handling of politically sensitive topics. While DeepSeek is programmed to align with government narratives, Manus provides factual, uncensored responses.

For example, when asked about the 1989 Tiananmen Square crackdown, it described the event accurately, stating that “the Chinese government carried out a violent crackdown on pro-democracy protesters”, followed by a detailed six-paragraph explanation.

Similarly, when questioned about human rights issues in Xinjiang, the AI gave a comprehensive and uncensored response.

Manus itself stated that it does not “intentionally censor factual information”, explaining that its goal is to offer “balanced, objective information to help users understand complex subjects.”

Tech analyst Li Jianggan from Momentum Works suggested that this could be because Manus is task-focused, and its developers may not have implemented strict content controls like those found in DeepSeek and ChatGPT.

Can Manus Be the Next DeepSeek?

Experts remain divided on whether the new AI design can match the success of DeepSeek. While its advanced enterprise automation capabilities make it a potential game-changer, it faces challenges in scaling up, ensuring stability, and meeting regulatory requirements.

According to Zhang, Manus could become a major player in enterprise AI if it successfully addresses these concerns. However, Li Jianggan cautioned that the system still has room for improvement.

Since Manus is designed to handle real-world tasks, its development will need to overcome complexity and unpredictability. Whether it can achieve mainstream success remains a “big question mark.”

With inputs from Reuters