New Breakthrough in AI Development: Manus Model Surpasses Other Models of the Same Level, Raising Security Concerns

Recently, the Manus model achieved breakthrough results in the GAIA benchmark test, outperforming other large language models of the same tier. This achievement demonstrates Manus's exceptional ability to handle complex tasks, such as multinational business negotiations that involve multiple skills. The advantages of Manus primarily lie in dynamic goal decomposition, cross-modal reasoning, and memory-enhanced learning. It can break down large tasks into hundreds of subtasks while processing various types of data and continuously improve decision-making efficiency and reduce error rates through reinforcement learning.

This development has once again sparked discussions within the industry about the path of AI development: should it pursue a single intelligence route towards Artificial General Intelligence (AGI), or a distributed route of multiple agent systems (MAS) working in collaboration? Both paths have their pros and cons. The AGI route aims to approach human-level comprehensive decision-making capabilities with a single system, while the MAS route focuses on coordinating multiple specialized intelligent agents to work together.

However, as AI systems become increasingly intelligent, their potential risks are also on the rise. The main concerns include:

1. Data privacy: In sensitive areas such as healthcare and finance, AI systems need to access a large amount of private data.
2. Algorithmic Bias: AI may exhibit unfair or discriminatory tendencies in certain decisions.
3. Security vulnerabilities: Highly intelligent systems may become important targets for hacker attacks.

To address these challenges, the industry is exploring various encryption technologies and security models:

• Zero Trust Security Model: Emphasizes strict validation of every access request.
• Decentralized Identity (DID): Achieve verifiable and persistent identity recognition without centralized registration.
• Fully Homomorphic Encryption (FHE): Allows computation on encrypted data, protecting privacy.

Among them, fully homomorphic encryption is considered one of the key technologies to solve security issues in the AI era. It can protect user privacy at the data level, achieve encrypted model training at the algorithm level, and use threshold encryption to protect communication at the collaborative level.

Despite the fact that security technology has always been a hot topic in the cryptocurrency field, many innovative projects have not received sufficient attention. For example, early decentralized identity projects and blockchain networks adopting a zero-trust model have failed to maintain long-term popularity in the market. Currently, some emerging FHE projects are attempting to apply this technology to practical scenarios and are collaborating with several tech giants.

As AI technology continues to approach human intelligence levels, establishing a robust security defense system becomes increasingly important. Technologies such as fully homomorphic encryption can not only address current security challenges but will also lay the groundwork for the future era of strong AI. On the path to AGI, these security technologies are no longer optional but are essential for ensuring the reliable operation of AI systems.