Enabling Agentic AI in the financial system requires zero-knowledge proof and privacy protection technology

Source: Chainlink Oracle
Original link: https://mp.weixin.qq.com/s/-mnpNrJP17RK-q1CrlQk9Q

---

The financial market is entering a new era with the emergence of "Agentic AI". This is a new model where autonomous, specialized Agents reason, act, and collaborate to address complex multi-step challenges. These Agents operate independently but excel when interacting with each other, leveraging their unique capabilities to efficiently achieve common goals, such as conducting compliance checks, creating comprehensive credit memorandums, or even simulating the impact of monetary policies.

Based on my work experience with AI systems, the most important breakthroughs often come not from individual systems, but from their ability to collaborate. The shift towards Agentic AI reflects a major trend in AI development: moving from isolated tools to interconnected systems designed to work seamlessly together. These systems are distinguished by their ability to integrate the following characteristics:

• Perception: Agents collect and interpret data relevant to their specialized domain through direct interaction with their environment, user input, or querying external systems.

• Reasoning and Planning: Agents analyze contextual information, evaluate goals, and develop strategies and plans that combine autonomous decision-making with human-aligned objectives.

• Tool Usage and Collaboration: Agents interact with other Agents and external systems, effectively executing tasks by utilizing tools and shared resources. These tools can be custom-developed for Agents or external systems such as enterprise applications or internet-hosted services.

• Execution: Agents take action based on their decisions, coordinating workflows to achieve results that sometimes require collective effort from specialized participants.

This set of capabilities enables Agents to adapt to changing conditions, collaborate seamlessly, and autonomously execute complex tasks, especially those that cannot be simplified into linear workflows. As such, they provide an alternative to rigid, rule-based systems that typically break down or require human intervention when faced with unexpected situations. In contrast, Agents can adapt to these situations in a non-deterministic manner.

While the autonomy and adaptability of Agentic AI can bring immense value in industries like finance or the public sector, this autonomy also brings risks and challenges. For example, how do they establish trust with other Agents and various stakeholders in these systems when Agents play roles previously performed by humans? How can we ensure their decisions are based on reliable real-world data when Agents make decisions? These are some of the issues we must address to unlock the full potential of Agentic AI.

[The rest of the translation follows the same professional and accurate approach, maintaining the original structure and meaning while translating to English.]

Zero-knowledge proofs (ZKPs) provide a cryptographic solution to the privacy paradox, allowing one party (an Agent) to prove the validity of a statement without revealing any additional information. While ZKPs are widely used in DeFi and the Web3 world, they can also play a crucial role in establishing trust in decentralized, multi-Agent systems. ZKPs' main advantages for Agentic AI: 1. Trust between Agents: Agents can securely verify each other's outputs, ensuring reliable collaboration without revealing unnecessary details. For example, when an Agent performs a task within its organizational boundaries and then passes the output to another Agent, it also provides a zero-knowledge proof demonstrating that the task was completed according to the organization's standards and requirements. 2. Credentials Verification without Disclosure: Agents can prove compliance (such as regulatory adherence) without exposing sensitive data and can also demonstrate proper authorization from their owners. 3. Minimizing Attack Surface: ZKPs limit data exposure, reducing vulnerabilities and enhancing security. 4. Reliable Decision-Making: Agents can verify the authenticity of external data, ensuring decisions are based on trustworthy information. For example, Agents can use data from decentralized oracle networks or Chainlink Data Feeds to provide critical real-world context for their decisions and actions. By leveraging ZKPs, Agentic AI can achieve secure, efficient, and private collaboration even in environments with initially low trust levels. Agentic AI in Financial Applications Agentic AI is poised to reshape the financial industry by automating complex processes, enhancing risk management, and improving decision-making. Its ability to deploy professional Agents that can work autonomously and collaborate with each other and humans releases new efficiencies and capabilities across various applications. [The rest of the translation follows the same approach, maintaining the specified translations for technical terms and preserving the structure of the original text.]