To Defend Against Cyber Attacks, We Must Teach AI Agents to Hack

Чтобы защититься от кибератак, мы должны научить ИИ-агентов взламывать системы
Terry Yue Zhuo, Yangruibo Ding, Wenbo Guo, Ruijie Meng
2026-02-01

AI cyber attacksautomated vulnerability discoverycyber range governanceoffensive AI capabilitiesoffensive security intelligence
For over a decade, cybersecurity has relied on human labor scarcity to limit attackers to high-value targets manually or generic automated attacks at scale. Building sophisticated exploits requires deep expertise and manual effort, leading defenders to assume adversaries cannot afford tailored attacks at scale. AI agents break this balance by automating vulnerability discovery and exploitation across thousands of targets, needing only small success rates to remain profitable. Current developers focus on preventing misuse through data filtering, safety alignment, and output guardrails. Such protections fail against adversaries who control open-weight models, bypass safety controls, or develop offensive capabilities independently. We argue that AI-agent-driven cyber attacks are inevitable, requiring a fundamental shift in defensive strategy. In this position paper, we identify why existing defenses cannot stop adaptive adversaries and demonstrate that defenders must develop offensive security intelligence. We propose three actions for building frontier offensive AI capabilities responsibly. First, construct comprehensive benchmarks covering the full attack lifecycle. Second, advance from workflow-based to trained agents for discovering in-wild vulnerabilities at scale. Third, implement governance restricting offensive agents to audited cyber ranges, staging release by capability tier, and distilling findings into safe defensive-only agents. We strongly recommend treating offensive AI capabilities as essential defensive infrastructure, as containing cybersecurity risks requires mastering them in controlled settings before adversaries do.
1
AI agents are expected to automate vulnerability discovery and exploitation across thousands of targets, undermining the assumption that tailored cyberattacks are too costly to scale.
2
AI-agent-driven cyberattacks are characterized as inevitable, requiring defenders to shift from solely preventive safeguards toward offensive security intelligence.
3
Data filtering, safety alignment, and output guardrails cannot reliably prevent misuse by adversaries controlling open-weight models or developing offensive capabilities independently.
4
Responsible development should restrict offensive agents to audited cyber ranges, stage deployment by capability tier, and distill results into defensive-only agents.
5
The paper recommends comprehensive attack-lifecycle benchmarks and trained agents capable of discovering real-world vulnerabilities at scale.

AI-agent-driven cyber attacks and offensive AI capabilities for automated vulnerability discovery and exploitation

The scalability, inevitability, defensive implications, and responsible governance of AI-agent-driven vulnerability discovery and exploitation across large numbers of targets

Publication Details
Publication Date
2026-02-01
Journal
Publisher
ISSN
Access Type
Author Information
Authors
Terry Yue Zhuo
Yangruibo Ding
Wenbo Guo
Ruijie Meng
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%