AI at the Gates: Combating AI-Driven Assaults on the Customer Experience
summary
A real marketplace attack uses task-driven AI agents that mimic clicks, keystrokes and swipes to scale theft, causing a reported $9 million loss. Agents run as monitors, add-to-cart bots and checkout workers, use harvested cookies, residential proxies and real identity data, then add payees and express transfers just before weekends to evade manual review. Detection works by linking session IDs across network, product and fraud logs; effective defenses are honeypots, poisoned responses and offline analysis rather than immediate blocking.
AI-written summary · about
chapters (9)
- 0:00AI-driven attacks on CX
- 3:11Rise of AI agents in fraud
- 7:47Sneaker bots to AI tools
- 14:18AI agents evading detection
- 17:53Strategies against AI fraud
- 23:15Detect and mitigate attacks
- 27:52Honeypots in defense
- 31:23Security vs CX in fintech
- 35:54Q&A
related talks
11mHow to Navigate Emerging AI-Driven Security
40mAI As The New Hacker Frontier
13mAI-Powered Payment Systems
01h 01mExploiting the supply chain
57mAI as the New Hacker Frontier
11mAI-Powered Cybersecurity in the Age of Advanced Threats
49mGenAI and Cybersecurity: Safeguarding the Digital Age
55mAI Security in Practice: Protecting Your AI-Powered Applications