AI-Driven Cybersecurity Challenges and Solutions in Web3
- Anti-money laundering penalties exceeded $900 million in the first half of 2025 due to intensified machine-speed attacks.
- Agentic AI systems are evolving from supportive roles to autonomous operational actors, executing end-to-end incident responses with minimal human intervention.
- Web3 protocols will adopt real-time, onchain circuit breakers and mandatory audits for AI logs to enhance security measures.
- Certik emphasizes the need for explicit roles and bounded authority for AI agents, maintaining human supervisory accountability due to retained liability.
- Natalie Newson from Certik highlights defenders’ home-field advantage, as they can preemptively analyze protocol behavior and monitor public ledger activities.
The rise of agentic AI in cybersecurity is shifting roles from support tools to autonomous actors capable of executing complex tasks independently. Despite this evolution, organizations must ensure these systems operate under strict governance frameworks to mitigate systemic risks.
With anti-money laundering penalties surpassing $900 million in early H1 of 2025, the push towards integrating real-time monitoring and audits is essential for countering machine-speed attacks effectively. (Source)
