The platform reported that in the first half of the year, AI rejected more than 30,000 suspicious withdrawal requests, protecting approximately 20,000 accounts. The average duration of request checks was 4.7 minutes.
Bybit states that AI is primarily used to increase the speed of threat detection, while responsibility for making complex decisions still lies with security specialists.
“The cybersecurity arms race has entered an era of minutes. Using AI to enhance our security and risk management capabilities, as well as ensuring the security of the AI systems themselves, is our top priority, while human judgment remains at the center of critical security decision-making,” said Bybit representative David Zong.
The crypto exchange also identified $212 million in cryptocurrency potentially linked to fraud and blacklisted more than 10,000 malicious addresses. To do this, it used behavioral analysis on the blockchain and AI-powered monitoring to identify new fraud patterns.
The company notes that artificial intelligence shortens the cycle of defensive operations. Bybit’s AI-based security audit identified vulnerabilities 3–5 times faster than manual checks. Some operations are now executed in hours, whereas previously people spent weeks on them.
