Coinbase has rebuilt its anti-fraud stack by tightly integrating machine learning models with a fast rules engine, reducing response times to new scam patterns from days to hours, just as TRM Labs warns that crypto fraud is now an AI-supercharged industry costing tens of billions per year.
Coinbase has upgraded its anti-fraud stack by tightly integrating machine learning models with a rules engine, reducing response time to new fraud patterns from days to just a few hours, as AI-enabled scams increase in the crypto sector.
The company describes a two-pronged strategy where “models [are] responsible for long-term defense, rules [are] responsible for a rapid response,” all housed in a unified framework that allows rules to capture new types of fraud that can then be fed back into models to strengthen overall defenses over time.
Coinbase says it has turned what was previously a manual and slow rule creation workflow into a target=”_blank” href=”https://www.rootdata.com/news/618934″>
Coinbase’s new fraud playbook
As part of the overhaul, rule backtesting performance has been improved by more than tenfold, allowing Coinbase to trial and ship new protections much faster as scam behavior evolves in real time.
According to Coinbase, the system now uses machine learning to recommend control parameters, with the goal of “reducing false positives, combating fraud, and minimizing the impact on regular users,” an important balance for a major exchange that handles billions in trading volume.
The latest upgrade builds on previous efforts detailed in a Coinbase blog on advanced machine learning models, in which the company said its mission is “to continue building scalable, adaptive, blockchain-aware ML systems that enable Coinbase to effectively manage risk to its products” without degrading the user experience.
AI arms race against crypto fraud
This move comes as crypto fraud has industrialized.
Blockchain intelligence firm TRM Labs reported that global crypto fraud would amount to approximately $35 billion by 2025, and warned that when underreporting is included, “total annual losses are likely to exceed $200 billion worldwide.”
In a separate 2026 crime report, TRM said illicit crypto flows will reach a record $158 billion by 2025, with scam networks increasingly resembling professional firms and AI tools accelerating mass impersonation and reach.
Coinbase’s own Chief Information Security Officer, Philip Martin Lunglhofer, has previously said that the exchange is seeing more and more “AI use cases to detect fraud” and is already using machine learning to monitor user activity and support chats for signs of scams or account takeovers.
The exchange’s latest investment in automated, event-driven rule generation and potential ‘one-click conversion’ of efficient rules into model functions aims to bring Coinbase closer to a fully automated risk management system, as fraudsters themselves weaponize AI to investigate and exploit weaknesses faster than ever.
For broader context on Coinbase’s security posture and user protection efforts, readers can refer to Coinbase’s fraud-focused blog posts on machine learning and compliance, as well as past coverage of Coinbase scam activities and crypto fraud trends on crypto.news.