Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
FollowingThe changing role of AI agents in cybersecurity and financial crime investigations
According to a new report released by blockchain security company Certik, artificial intelligence (AI) agents are transcending the role of assistive tools in cybersecurity and financial crime investigations. The report points out that the new generation of systems is capable of reasoning, using tools, gathering evidence, acting in a real-time environment, and reviewing results with little human intervention.
For many years, machine learning systems have largely supported the work of analysts by flagging abnormal logins, scoring transactions, and preparing reports. Certik said newer agents could perform a wider range of tasks, changing their role from "identifying problems" to "assisting in executing responses."
Agents take on more security tasks
In the Security Operations Center, an agent can investigate suspicious login events by examining device records, location data, and threat intelligence sources, and then take action. For example, it can suspend an account and record its steps for human review.
Similarly, such systems have emerged in the Web3 security space, including scenarios such as contract classification, transaction risk scoring, and tracking stolen funds. As agents handle routine investigations, it is expected that human experts will focus more on supervision.
One reason for this shift is the extremely fast speed of some attacks. Lightning loan vulnerabilities can drain protocol liquidity in seconds, while stolen cryptocurrency can move across bridges and mixers in hours. This leaves very limited time for the human team to investigate and respond.
In addition, the shortage of cybersecurity and compliance professionals is also an important factor, while regulatory pressure has increased the workload. The security company cited more than $900 million in AML (anti-money laundering) fines for the first half of 2025 to show the consequences companies can face when controls fail.
AI autonomy brings new risks
Certik believes that companies should view autonomous agents as workforce participants rather than ordinary software. Their actions do not absolve humans of responsibility, so the companies deploying these agents remain responsible for the end result.
This accountability becomes increasingly important as agents take on more tasks in financial and crypto operations. In the financial world, agents can handle anti-money laundering tasks; in the crypto world, they can detect vulnerabilities, track funds and monitor transactions.
However, the report also highlights the risks posed by erroneous output, reduced human censorship, and deliberate attacks on AI systems. Certik recommends keeping records of inputs and operations, clearly limiting the scope of autonomous decision-making, conducting adversarial testing, and appointing a person in charge for each agent.
As AI systems have greater authority over security and compliance tasks, enterprises need controls that match their extended role. Certik said that without strong oversight, automation could replace familiar problems with more difficult to explain failures.
