Crypto exchange Bybit says its artificial intelligence-powered monitoring systems have helped prevent approximately $300 million in potential user losses, highlighting the growing role of automated risk detection as fraud attempts escalate across the digital asset sector.

According to the company, its AI-driven Dynamic Risk-Based Protection System identified roughly $500 million in suspicious withdrawal attempts during the fourth quarter of 2025. Of that amount, around $300 million was successfully blocked or recovered before funds could be transferred to fraudulent addresses. Bybit indicated that more than 4,000 user accounts were protected during the period.

The exchange said the surge in flagged transactions reflects a broader increase in impersonation scams, credential theft, and coordinated social engineering campaigns targeting crypto users. Industry data has shown that digital asset fraud continues to evolve in sophistication, with attackers leveraging automation and AI-generated tactics to bypass traditional safeguards.

Layered monitoring and real-time intervention

Bybit’s security framework operates through a tiered response model designed to assess risk severity in real time. Low-risk anomalies, such as unusual withdrawal timing or unfamiliar device logins, trigger automated prompts or additional verification checks. Medium-risk activity can prompt immediate on-platform alerts urging users to confirm transaction details. High-risk signals, including transfers to wallets linked to known scam networks, may result in instant withdrawal blocks and temporary cooling-off periods.

The company said its proprietary AI models analyze behavioral data, transaction patterns, and on-chain intelligence to detect potentially fraudulent activity before funds leave the platform. By combining internal monitoring with blockchain analytics, the system aims to identify red flags that might not be apparent through manual review alone.

Security executives at the exchange emphasized that the objective is not to indiscriminately freeze accounts, but to balance user protection with operational continuity. The cooling-off mechanism, for example, is designed to provide users with time to reassess transactions flagged as high risk without permanently restricting access.

Rising industry focus on proactive defense

The announcement comes as crypto platforms face mounting pressure to strengthen safeguards against increasingly complex scams. Analysts note that fraudsters have adopted automation tools and AI-driven impersonation tactics to scale attacks, making reactive security approaches less effective.

As a result, exchanges are investing more heavily in predictive monitoring and automated threat detection. AI-based systems can process large volumes of transactional data at speeds unattainable through manual oversight, allowing platforms to intervene before losses occur rather than attempting recovery afterward.

The broader digital asset industry has reported billions of dollars in annual losses tied to scams and exploit schemes. While centralized exchanges are not immune to these risks, many have begun integrating advanced analytics and machine learning frameworks to reduce exposure.

For Bybit, the reported interception of $300 million in suspected fraudulent withdrawals represents both a risk management milestone and a public demonstration of its security infrastructure. Market observers note that trust and asset protection remain central to exchange competitiveness, particularly as institutional participation in crypto markets expands.

As cyber threats continue to evolve alongside digital asset adoption, AI-driven monitoring systems are likely to become standard components of exchange security architecture. The effectiveness of these tools may play an increasingly important role in shaping user confidence and regulatory perceptions of the crypto trading environment.