Postmortems can’t stop AI-powered crypto fraud

AI is dramatically escalating crypto fraud, with scams utilizing deepfakes and voice clones, leading to billions stolen, as seen by the $2.17 billion lost in the first half of 2025 alone. The industry's current response—relying on reactive tools like audits, blacklists, and post-incident write-ups—is too slow and ill-suited for machine-speed threats. The author argues that the focus must shift from reactive defense to proactive, embedded resilience within transaction logic. This involves developing wallets that can detect anomalies in real-time, intervene before harm occurs (e.g., by requiring extra confirmations or analyzing transaction intent), and utilizing shared intelligence networks among security providers. If the crypto industry fails to voluntarily adopt these systemic protections, regulators will likely impose rigid frameworks. Ultimately, the goal is to make irreversible loss intolerable by embedding trust and resilience into the transaction fabric, rather than relying solely on user education or retrospective analysis.

(Source：Cointelegraph)