Stories
Interpol Warns: AI Fuels Faster, Bigger Cybercrime Threat to Corporations
- •Interpol's Bjorn R Watne states AI accelerates existing cybercrime tactics like scams and fraud, making them faster and easier to scale, rather than creating entirely new methods.
- •AI enables scam operators to target multiple victims simultaneously, with improved translation tools and digital identities making fraudulent interactions harder to detect.
- •Watne advises companies to identify their "crown jewels" – critical assets – and use threat intelligence to tailor defenses against potential adversaries.
- •He highlights a disconnect where some senior managements haven't recognized that all companies are now essentially IT companies, underscoring the need for robust cybersecurity strategies.
- •Watne expresses concern about agentic AI, where systems act on users' behalf, noting that errors could have significant consequences, especially when AI interacts with the physical world.
Key Questions Answered
Interpol global chief information security officer Bjorn R Watne stated that artificial intelligence is an evolution rather than a revolution, primarily making existing tactics like scams and fraud faster and easier to scale. AI enables scam operators to target multiple victims simultaneously while using improved translation tools and digital identities to make fraud harder to detect.
Companies should first identify their crown jewels, which are the most critical assets needed to keep operations running, and determine who might want to disrupt or steal them. Businesses can then utilize threat intelligence to understand the tactics of potential adversaries and tailor their defenses accordingly rather than trying to protect against everything at once.
Bjorn R Watne expressed particular concern regarding agentic AI as these systems gain the capability to act directly on users' behalf. Errors in such systems could lead to consequences that go beyond providing incorrect information, particularly when AI is deployed in the physical world.
