Crypto for Advisors: Strengthening defenses against AI fraud

Happy Thursday, advisors!

In today’s newsletter, Kriti Bansal examines the rise of AI-driven fraud and provides a framework of rigorous financial controls to help advisors secure client assets against sophisticated impersonation tactics.

Then, in “Ask an Expert,” Varun Choudhary, CEO of ORO, discusses how money managers can automate defensive layers against fraud by transitioning to programmable smart accounts and using automated monitoring to create programmatic security guardrails at the account level.

Happy reading.

AI Is changing crypto fraud — the best defense is old-fashioned financial control

As AI makes impersonation cheap and convincing, an advisor’s strongest protection isn’t a better eye for fakes — it’s the verification, separation of duties and reconciliation they already know.

For most of crypto’s history, fraud was a volume business: send enough phishing messages and a few would land. But artificial intelligence has changed the economics of fraud. Deception is now cheaper to produce, more personalized and markedly more convincing — and increasingly, it arrives in the form of someone your client already trusts.

The scale is significant: the FBI’s Internet Crime Complaint Center reported a record $20.9 billion in cybercrime losses in 2025, with cryptocurrency the most common payment channel (FBI). Chainalysis estimates that as much as $17 billion flowed to crypto scams over the same period, and found that operations linked to AI tools were roughly 4.5 times more profitable than those without. The average scam payment more than tripled year over year, to $2,764.

For advisors, the instinct is to get better at spotting fakes. That is unlikely to be a durable strategy: synthetic video and cloned audio are already convincing and continue to improve. The more reliable protection is one advisors already understand: financial control. An advisor’s fiduciary duty to safeguard client assets, and the SEC’s custody rule under the Investment Advisers Act, do not depend on detecting a deepfake. They depend on verification, separation of duties and reconciliation. In digital assets, where a transaction is irreversible once settled, those controls matter more, not less.

• Impersonation at scale. Chainalysis recorded a roughly 1,400% increase in impersonation scams. Real-time face-swap tools, voice cloning and large language models let a bad actor appear as a client’s advisor, a fund principal or a support agent — including on live video. Verifying identity by “hopping on a call” is no longer sufficient.
• Automated persistence. “Pig butchering” investment scams, built on weeks of relationship-building, cost victims $7.2 billion in 2025. AI systems now sustain those conversations continuously and across many targets at once.
• Low-cost fabrication. Convincing fake trading platforms, synthetic testimonials and fabricated news segments can now be produced in minutes rather than by a professional team.