The landscape of cryptocurrency security is undergoing a fundamental transformation, shifting from a reactive battle against isolated hacks to a proactive, automated arms race driven by artificial intelligence. According to Simone Maini, CEO of blockchain analytics firm Elliptic, the single biggest emerging risk is not a record-breaking heist but the sheer velocity and volume of AI-driven financial activity that could overwhelm traditional compliance systems designed for human-paced markets.
In a recent interview, Maini outlined how AI agents—autonomous software programs capable of executing complex financial transactions without human intervention—are being deployed by both bad actors and legitimate financial institutions. This duality is creating an environment where the speed of operations far outpaces the ability of human analysts to monitor and respond. “We are entering an era where machines trade and transact at millisecond speeds, and our compliance infrastructure was built for a world where humans make decisions over minutes or hours,” Maini said.
The Rise of AI Agents in Crypto
AI agents have already begun to infiltrate the cryptocurrency ecosystem. These programs can manage wallets, execute trades, interact with decentralized finance protocols, and even simulate human behavior to bypass know-your-customer checks. For malicious actors, the barrier to entry has lowered dramatically. Hacking tools that once required specialized programming skills are now available as AI-powered services, allowing even novice criminals to launch sophisticated attacks at scale.
One growing concern is the use of AI to automate phishing scams and social engineering attacks. Instead of sending thousands of identical emails, AI agents can personalize messages in real time, scraping social media data to craft convincing appeals targeting specific individuals or organizations. The result is a dramatic increase in both the volume and success rate of these scams, often leaving compliance teams scrambling to identify patterns before significant damage occurs.
Moreover, AI agents are increasingly used to manipulate decentralized markets. They can engage in front-running, sandwich attacks, and flash loan exploits with a speed and precision that human traders cannot match. These activities not only harm individual investors but also erode trust in the entire decentralized finance ecosystem. As Maini noted, “The problem is not that AI will enable a single $1 billion hack, but that millions of small, automated frauds will collectively dwarf any single event.”
The Compliance Challenge
Traditional crypto compliance relies on blockchain analytics tools that flag suspicious transactions based on pre-defined rules and known addresses. These systems are effective when illicit activity occurs at a human scale—where a few thousand transactions per day need to be reviewed. However, AI agents can generate millions of transactions per hour, each appearing innocuous on its own but collectively forming a pattern of money laundering or fraud.
Monitoring systems built for human-paced markets simply cannot keep up. Risk analysts would need to triage alerts in real time, but the sheer number of false positives and the complexity of AI-generated patterns make manual review impractical. As a result, many suspicious activities escape detection until it is too late. “Compliance teams are drowning in alerts, and the quality of those alerts is deteriorating because AI is learning to evade detection,” said Maini.
The problem is compounded by the global nature of cryptocurrency. Different jurisdictions have varying regulations, and an AI agent can instantly route funds through multiple countries, exploiting regulatory loopholes faster than authorities can close them. This creates a legal and operational headache for compliance officers who must navigate a patchwork of rules while trying to keep pace with automated criminals.
Elliptic’s Response: AI Agents for Good
To counter these threats, Elliptic is developing a new generation of compliance tools that leverage AI agents themselves. The idea is to fight fire with fire: deploy autonomous monitoring agents that can analyze blockchain data in real time, identify suspicious patterns, and even predict future attacks before they occur. These AI agents continuously learn from new data, adapting to evolving tactics employed by malicious actors.
This approach represents a significant shift from static rule-based systems to dynamic, machine learning-driven platforms. Elliptic’s agentic compliance system, which the company began rolling out in early 2026, uses neural networks to model normal transaction behavior and flag anomalies. It can also cross-reference data from multiple blockchains, decentralized exchanges, and off-chain sources to build a holistic picture of risk associated with any address or transaction.
“We are building a system that can think like an attacker but act as a defender,” Maini explained. “Our agents simulate potential attack vectors and automatically update risk scores, blocking suspicious transactions before they are executed.” This proactive defense is critical in an environment where speed is paramount. By automating the detection and response process, Elliptic aims to reduce the time between an attack initiation and its neutralization from hours to milliseconds.
The company’s efforts have attracted significant financial backing. In a recent funding round, Elliptic raised $120 million from investors including Nasdaq and Deutsche Bank. The capital will be used to expand its agentic compliance platform and hire additional talent in artificial intelligence and blockchain security. The involvement of major financial institutions underscores the growing recognition that conventional compliance methods are insufficient in an AI-driven world.
Broader Implications for the Crypto Ecosystem
The AI arms race in crypto compliance has implications far beyond individual firms. Entire regulatory frameworks may need to be overhauled to account for machine-paced finance. Currently, most anti-money laundering laws require human oversight of suspicious activity reporting, but if transactions occur in milliseconds, human review becomes impractical. Regulators may soon mandate the use of AI-powered compliance tools, just as they mandate transaction monitoring systems today.
Furthermore, the arms race is likely to accelerate as both sides deploy increasingly sophisticated models. The cost of developing AI agents is falling rapidly due to open-source libraries and cloud computing resources. This democratization of AI means that even small groups can field powerful agents capable of evading detection. In response, compliance firms must continually invest in research and development to stay ahead of the curve.
The evolution of crypto security also raises questions about privacy. AI agents require access to vast amounts of data to train their models, including transaction histories, wallet addresses, and behavioral patterns. Balancing the need for robust security with the right to pseudonymity—a core value of cryptocurrency—will be a delicate challenge. Overly aggressive surveillance could drive users to privacy-focused coins and decentralized exchanges that resist monitoring, further complicating the regulatory landscape.
Historical Context: From Hacks to AI
To understand the significance of this shift, one must consider the history of crypto security. A decade ago, the biggest threats were exchange hacks like the Mt. Gox collapse in 2014, where human error and poor security protocols led to the loss of 850,000 Bitcoin. The response was improved exchange security and the rise of custodial services. Then came smart contract exploits, such as the DAO hack in 2016, which prompted rigorous code audits and insurance products. In the early 2020s, ransomware and phishing attacks became prevalent, leading to the development of blockchain analytics tools that could trace stolen funds.
Each era brought new defenses, but the pace of change was manageable. Now, AI is compressing those cycles. Vulnerabilities that once took months to discover and exploit can now be found within days by automated scanning agents. The same technology that enables self-driving cars and personalized medicine is being weaponized against the financial system. As Maini puts it, “We've moved from a game of whack-a-mole to a game of cat and mouse, where the mouse has become as fast as the cat.”
The institutional adoption of cryptocurrency adds another layer of urgency. Banks, hedge funds, and pension funds are increasingly allocating capital to digital assets. They demand compliance solutions that meet traditional financial standards, but those standards were designed for a world without autonomous agents. To bridge the gap, firms like Elliptic are collaborating with regulators to define what “adequate” compliance looks like in an AI-enabled environment. Some experts predict that within five years, most large crypto transactions will be monitored by AI systems on both sides—the perpetrator and the defender.
The Road Ahead
Elliptic’s funding round is a bellwether for the industry. Nasdaq's participation signals that traditional market infrastructure providers view agentic compliance as essential for the legitimacy of crypto markets. Deutsche Bank’s involvement indicates that major lenders are preparing to offer custody and trading services that meet regulatory standards. Together, these partnerships could accelerate the adoption of AI-powered compliance across the entire financial sector, not just crypto.
As the technology matures, new best practices will emerge. Some organizations may pool data to train shared AI models, creating a collective defense network. Others may specialize in specific niches, such as AI agents designed to detect money laundering in decentralized finance or agents focusing on ransomware payments. The key will be interoperability: ensuring that different compliance agents can communicate and share threat intelligence without compromising sensitive information.
Despite the challenges, Maini remains optimistic about the potential of AI to improve security. “The same tools that make it easy to attack also make it possible to defend at scale,” she said. “We have the opportunity to build a financial system that is more resilient than anything we have today.” However, she warns that this opportunity comes with a responsibility to act quickly. The AI arms race in crypto security is not a distant threat—it is already underway, and the side that moves fastest will define the rules of engagement for years to come.
The transformation of Elliptic from a traditional analytics firm to an agentic compliance pioneer reflects a broader trend in cybersecurity. As AI continues to evolve, the line between human and machine oversight will blur. Compliance teams will no longer be made up solely of analysts and investigators; they will include data scientists, AI engineers, and risk modelers. The future of crypto security is autonomous, adaptive, and relentless—mirroring the very forces it seeks to contain.
Source: Coindesk News