AI Is Accelerating Fraud — But the Playbook Hasn’t Changed
TREE NEWS reports: INTERPOL’s global chief information security officer, Bjorn R. Watne, has a blunt message for companies racing to defend themselves against artificial intelligence: before you buy a single AI security tool, answer one question — what are your crown jewels? Watne made the remarks at Tech Week Singapore, warning that AI is not inventing new crimes so much as supercharging old ones, letting fraudsters run familiar scams at unprecedented speed and scale.
For the crypto industry, that warning lands with particular force. Digital assets have long been a favorite target for social engineering, phishing, fake exchange apps, romance scams and wallet-draining malware. AI now lets attackers generate flawless phishing emails in any language, clone voices to impersonate exchange support staff, and spin up convincing deepfake videos of founders and influencers promoting fraudulent tokens — all at near-zero marginal cost.
Why Crypto Is a Prime Target
Several structural features make crypto uniquely exposed:
- Irreversibility: On-chain transactions cannot be clawed back, so a single successful drain is permanent.
- Pseudonymity: Attackers can move funds across chains and mixers before victims even notice.
- Retail-heavy user base: Many holders lack the security hygiene of institutional traders.
- Trust-based interfaces: Wallets, dApps and token launches rely heavily on brand trust — exactly what deepfakes exploit.
Watne’s framework — identify the assets that would cause existential harm if compromised, then build defenses around them — maps cleanly onto crypto. For an exchange, that might be hot wallet private keys and withdrawal authorization systems. For a DeFi protocol, it could be admin keys, oracle feeds or upgradeable contract controls. For a custodian, it’s the signing infrastructure itself.
From Perimeter Defense to Asset-Centric Defense
The shift Watne describes is significant. For years, security teams focused on perimeter defense: firewalls, endpoint protection, employee training. But AI-driven attacks compress the time between reconnaissance and exploitation, making broad perimeters harder to hold. An asset-centric approach — knowing exactly what must never be compromised — allows firms to concentrate resources where failure is catastrophic rather than merely inconvenient.
This aligns with a broader trend in crypto security: multi-signature wallets, hardware security modules, time-locked withdrawals, and increasingly, AI-assisted anomaly detection that flags suspicious transaction patterns in real time. The irony is that the same AI arms race cuts both ways — defenders can use machine learning to detect fraud faster, but only if they know what they’re defending.
The Regulatory Overlay
Regulators are watching closely. As AI-generated fraud rises, expect heightened scrutiny on crypto firms’ cybersecurity disclosures, incident reporting timelines, and consumer protection obligations. Jurisdictions advancing licensing regimes are increasingly embedding security requirements into the approval process, making Watne’s question not just a strategic exercise but a compliance necessity.
What Comes Next
The takeaway for crypto firms is clear: AI will not create a new category of threat so much as industrialize the existing one. The winners will be those who can articulate, in plain terms, what they cannot afford to lose — and then engineer every layer of their stack around protecting it. In an industry built on the promise of self-custody and trustless systems, the oldest security question may prove the most important one.
