Zhou Hongyi Calls for AI-Native Cybersecurity Transformation
TREE NEWS reports: At the 4th Cyberspace Security (Tianjin) Forum, Zhou Hongyi, founder of 360 Group, delivered a keynote arguing that AI has fundamentally altered the landscape of cyber offense and defense. Traditional security measures are no longer sufficient, he said, urging the industry to adopt an automated attack-defense framework centered on ‘using models to govern models’ (以模治模).
Key Points from Zhou’s Address
- AI has changed the game: Both attackers and defenders now leverage AI, making conventional signature-based defenses obsolete.
- New defense paradigm: Zhou proposes an automated system that uses AI to detect vulnerabilities, orchestrate responses, and secure AI agents themselves.
- Comprehensive AI safety: Ensuring large language models (LLMs) are secure, their outputs are reliable, and their results are trustworthy.
- Lifecycle governance for agents: AI agents must be controlled throughout their lifecycle—auditable, intervenable, and stoppable.
Industry Analysis
Zhou’s remarks resonate with a growing recognition that AI is a double-edged sword. On one hand, AI-powered attacks can scan for vulnerabilities and craft phishing campaigns at scale. On the other, AI offers the only viable defense at machine speed. His call for ‘using models to govern models’ aligns with emerging practices in the cybersecurity industry, where AI-driven SOAR (Security Orchestration, Automation, and Response) platforms are gaining traction.
For the crypto and Web3 sector, this has direct implications. DeFi protocols and blockchain networks are increasingly reliant on AI for smart contract auditing, fraud detection, and governance. If AI systems themselves become attack vectors—through prompt injection, model poisoning, or adversarial examples—the entire decentralized finance ecosystem could face new risks. Zhou’s emphasis on agent lifecycle management is particularly relevant as autonomous AI agents begin to manage digital assets and execute transactions on-chain.
Forward-Looking Perspective
The cybersecurity industry is at an inflection point. Zhou’s proposal suggests a future where AI is not just a tool but the core of a self-evolving security infrastructure. We can expect to see more investment in AI-driven threat intelligence, automated penetration testing, and real-time response systems. For blockchain networks, integrating such AI defenses will be crucial to maintain trust and security as the industry scales.
Moreover, the concept of ‘auditable, intervenable, stoppable’ AI agents could serve as a blueprint for AI governance in decentralized ecosystems. Smart contract-based kill switches and on-chain monitoring of AI agents might become standard practice. As Zhou emphasizes, the industry must act quickly to form a unified AI security framework before malicious actors exploit the gap.
