AI Security issues are increasingly prominent, and fully homomorphic encryption may become the best solution.

Recently, an AI system named Manus achieved breakthrough results in the GAIA benchmark test, outperforming large language models of the same level. Manus demonstrated strong independent task handling capabilities, capable of managing complex multinational business negotiations, including contract clause analysis, strategy formulation, and proposal generation. Compared to traditional systems, Manus has clear advantages in dynamic goal decomposition, cross-modal reasoning, and memory-enhanced learning.

The emergence of Manus has once again sparked discussions within the industry about the development path of AI: should it move towards a unified model of Artificial General Intelligence (AGI) or a collaborative model of Multi-Agent Systems (MAS)? This debate actually reflects the balance issue between efficiency and security in AI development. As singular intelligent systems gradually approach AGI levels, the risk of opacity in their decision-making processes also increases. While multi-agent collaboration can disperse risks, it may miss critical decision-making opportunities due to communication delays.

The progress of Manus also highlights the potential risks in AI development, such as data privacy, algorithmic bias, and adversarial attacks. For example, in medical scenarios, AI systems need access to sensitive genomic data of patients; in financial negotiations, there may be undisclosed financial information of companies involved. Additionally, AI systems may exhibit bias against certain groups during the hiring process or make incorrect judgments about terms in legal document reviews for emerging industries. More seriously, hackers may implant specific audio signals, causing AI systems to make erroneous judgments during negotiations.

In the face of these challenges, security technologies in the Web3 field may provide solutions. Among them, fully homomorphic encryption (FHE) technology is regarded as a powerful tool to address security issues in the AI era. FHE allows computations on data in an encrypted state, which means that sensitive information can be processed without decryption.

At the data level, FHE can ensure that all information input by users (including biometrics, voice, etc.) is processed in an encrypted state, such that even the AI system itself cannot decrypt the original data. At the algorithm level, the "encryption model training" achieved by FHE allows even developers to not directly view the AI's decision-making process. In terms of multi-agent collaboration, using threshold encryption technology can prevent a single node from being compromised, leading to global data leakage.

Although Web3 security technology may not have a direct connection with ordinary users, its importance cannot be overlooked. In this challenging field, if protective measures are not actively taken, users may never be able to escape information security risks.

Currently, there are some projects exploring the Web3 security field. For example, some projects have made certain progress in decentralized identity (DID) and zero-trust security models. In the field of fully homomorphic encryption (FHE), a certain project has already launched on the mainnet and has collaborated with several well-known institutions.

As AI technology approaches human intelligence levels, non-traditional defense systems are becoming increasingly important. FHE not only addresses current security issues but also lays the foundation for a more powerful AI era in the future. On the road to AGI, FHE is no longer an option but a necessary condition for ensuring the safe development of AI.