New Chinese AI agent draws DeepSeek comparison
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Aïda Amer/Axios
Just days after its announcement, a new AI agent named Manus is winning expert acclaim while stoking concern over another AI advance rooted in Chinese research and development.
Driving the news: Manus AI, named after the Latin word for hand, is billed as "a general AI agent that turns your thoughts into actions."
- In a video posted late last week, Manus' creator describes it more than "just another chatbot or workflow. ... It's a completely autonomous agent."
- "We see it as the next paradigm of human-machine collaboration and potentially a glimpse into AGI," Manus AI chief scientist Yichao "Peak" Ji says in the video.
Ji's demo shows Manus handling three separate tasks, sorting through resumes, identifying correlations in various stocks and searching through New York real estate — rating nearby schools and assessing how much the user can realistically afford.
- Manus does its work in the cloud, he says, which allows users to shut their laptop while the AI is working — but also raises concerns about data security and privacy.
Why it matters: As with DeepSeek, the advent of Manus is alarming some U.S. observers who worry that China is catching up in what is often cast as a race for AI supremacy.
Yes, but: For now, Manus is in invitation-only private testing.
What they're saying: Manus has sparked a flurry of online discussion, with enthusiasm for its capabilities, critiques of its limitations, and warnings about its implications for privacy and security.
- "It looks like Manus AI presents itself as a Chinese company (with its team based in China) while maintaining a legal entity in Singapore," AI and privacy expert Luiza Jarovsky writes in a newsletter.
- "From a data protection perspective, the key questions are: Where are its servers located? Is there any corporate affiliation to China? Are there data transfers to China?"
The other side: Skeptics are already questioning whether Manus can live up to the claims being made for it.
- Some early testers, like TechCrunch's Kyle Wiggers, found its performance on various tasks disappointing.
The big picture: Agents with various levels of autonomy have been widely seen as the next big thing in AI and were already an industry buzzword long before Manus' arrival.
Between the lines: Manus arrives at a moment of growing concern over dangers autonomous AI agents might pose.
- In a recent paper, four researchers at Hugging Face make a forceful argument that "fully autonomous AI agents should not be developed."
- "As history demonstrates, even well-engineered autonomous systems can make catastrophic errors from trivial causes," the authors write. "While increased autonomy can offer genuine benefits in specific contexts, human judgment and contextual understanding remain essential, particularly for high-stakes decisions."
