The Next Frontier: Intelligent Agents Take Center Stage

A significant shift is underway in artificial intelligence. As noted by Zhou Hongyi, founder of 360, during his address at the 2026 World Artificial Intelligence Conference (WAIC), the industry is transitioning beyond the era of large language models into the age of intelligent agents.

This evolution marks a fundamental change in capability. AI is no longer just a sophisticated tool for answering questions; it's becoming an autonomous entity capable of planning and executing complex tasks independently. From managing customer service workflows to analyzing financial data, these agents are beginning to operate with a level of autonomy previously unseen.

The Security Paradox of Rapid Deployment

While this autonomy accelerates real-world application and adoption, it introduces a new layer of complexity and risk. Zhou Hongyi highlighted that greater capability brings greater responsibility—and more significant security challenges.

When an AI system moves from offering suggestions to taking direct action in critical environments—be it financial trading, industrial control, or logistics—the potential impact of a flaw, bias, or malicious compromise grows exponentially. A single erroneous decision by an autonomous agent could have tangible, costly consequences.

The Imperative of Trustworthy AI

In response to these emerging threats, Zhou emphasized four non-negotiable pillars for future AI development: safety, trustworthiness, reliability, and controllability. These principles are critical for ensuring AI systems can be integrated into society responsibly.

  • Safety: Protecting agents from attacks, misuse, or causing unintended harm.
  • Trustworthiness: Ensuring decision-making processes are transparent and outcomes are aligned with human values.
  • Reliability: Maintaining consistent and stable performance in task execution.
  • Controllability： Preserving the ability for human oversight, intervention, and ultimate authority.

He noted that events like WAIC play a crucial role beyond showcasing innovation; they provide a vital forum for establishing global norms and governance frameworks for this new technological phase.

The New Competitive Landscape

The coming phase of AI competition will therefore hinge not just on raw technical prowess, but on the ability to build and deploy agents that are inherently secure and trustworthy. Organizations that prioritize embedding these principles into their development lifecycle from the outset will likely gain a significant advantage in both commercial adoption and scalable implementation. For developers and enterprises alike, security is shifting from an afterthought to a core design requirement.