A new Chinese artificial intelligence (AI) platform, Manus AI, has been making headlines globally, claiming to have developed the world’s first fully autonomous AI agent. Announced on March 5 by the Chinese start-up Butterfly Effect, Manus is being heralded as a potential game-changer in AI technology, following the success of another Chinese AI start-up, DeepSeek.
Dubbed “the first general AI agent,” Manus AI is designed to autonomously execute complex tasks without needing constant human input, a significant leap in AI capabilities. However, while the platform has garnered attention for its impressive performance, it has also raised alarms, particularly about potential data privacy risks.
What Can Manus AI Do?
Manus AI is not a typical chatbot or task automation platform. According to the company, Manus is a general AI agent that can autonomously perform a wide variety of tasks, ranging from property buying, programming video games, and analyzing stocks to planning travel itineraries.
The platform’s ability to perform these tasks goes beyond just responding to human commands; it autonomously executes the tasks with minimal human guidance, making it stand out in a crowded AI market. The AI is designed to turn the user’s thoughts into actions, with the company drawing a parallel to the concept of artificial general intelligence (AGI), where the AI mimics or surpasses human cognitive abilities.
“We see it as the next paradigm of human-machine collaboration and potentially a glimpse into AGI,” said Yichao “Peak” Ji, Manus AI’s chief scientist. While its potential is significant, experts remain cautious about its real-world applicability and the implications of such powerful AI agents.
How Does Manus AI Work?
Manus AI was not created from the ground up but instead leverages existing large language models (LLMs) such as Anthropic’s Claude and Alibaba’s Qwen. However, it has garnered significant attention due to its strong benchmark performance, claiming to surpass OpenAI’s Deep Research agent powered by the o3 model.
The platform’s autonomy allows it to execute complex tasks independently, marking a key step forward in the race for fully autonomous AI agents. It provides an intriguing glimpse into the future of AI-driven automation, with the promise of taking human-machine collaboration to new levels.
Expert Opinions: The Hype vs. Reality
Despite Manus AI’s ambitious claims, reactions from the AI community have been mixed. Some experts, such as Rowan Cheung, founder of The Rundown AI newsletter, have described Manus as “China’s second DeepSeek moment,” likening it to a viral sensation in the AI space. Cheung further described it as a combination of Deep Research, Operator, and Claude Computer, calling it “REALLY good.”
However, other experts remain skeptical. Bradford Levy, an assistant professor at the University of Chicago Booth School of Business, tested Manus and pointed out several shortcomings. According to Levy, the platform often fails to get things right on the first try and is prone to getting stuck in infinite loops. He also highlighted a lack of transparency around Manus’ capabilities, raising concerns about whether it can truly live up to the hype.
“There’s a good chance it won’t meet expectations,” Levy said, urging caution before fully embracing the platform. He also warned users about the risks of feeding sensitive data to AI platforms, especially considering Manus’ ties to China.
Concerns Over Data Privacy and Transparency
As Manus AI begins to gain traction, experts are also raising concerns about data privacy. Given the platform’s ties to China and its potential access to vast amounts of personal data, there is heightened scrutiny about how this data might be used, particularly in light of the country’s strict data privacy laws and its relationship with the Chinese Communist Party.
Levy’s warnings about transparency are echoed by others who fear that users may unknowingly expose themselves to privacy risks by relying on an AI platform of this nature. As Manus AI’s influence grows, the debate over the ethical implications of AI-powered platforms, especially those with ties to state-controlled entities, will likely intensify.
