Major Exchange's AI blocks $300M crypto scams

TLDR

Bybit says its AI security system blocked around $300 million worth of attempted crypto scams in 2025, mostly in Q4.

1. Bybit reports its AI-driven risk framework intercepted roughly $300 million in impersonation, scam, and fraud attempts in 2025.
2. The move shows how large exchanges are increasingly using AI to fight account takeovers, social engineering, and off-platform pig-butchering style scams.
3. Users still need to harden their own security, and a key thing to watch is whether other major exchanges match or exceed these AI protections.

Deep Dive

1. What Actually Happened

Bybit publicly highlighted a 2025 security milestone, stating that its AI-driven risk framework intercepted about $300 million in impersonation, scams, and fraud attempts, with a focus on Q4 2025 activity, according to a Chainwire press release distributed via TradingView and other outlets.

The framework is described as AI based, meaning it uses machine learning models to score risk in real time across logins, withdrawals, and other behaviors, then flags or blocks suspicious activity automatically.

This is prevention, not recovery: the claim is that those assets were stopped from being stolen before leaving user control or the platform, rather than clawed back after the fact.

2. Why AI Security On Exchanges Matters

Crypto scams increasingly rely on social engineering, fake trading platforms, and account compromise rather than purely technical exploits. Law enforcement reports and exchange statements show billions in losses across pig-butchering, impersonation, and romance-investment scams.

An AI risk engine lets an exchange connect many weak signals (IP changes, device fingerprint, message patterns, transaction graph) into a strong risk score, which is hard to replicate with manual reviews alone.

If Bybit can reliably stop hundreds of millions of dollars in fraudulent flows, it reduces direct user losses and also makes the platform less attractive to organized scammers who rely on easy off-ramps.

3. What Users Should Watch Next

First, see whether other top exchanges publish comparable security stats or explain similar AI frameworks. Competitive pressure could raise the minimum standard for risk controls across the industry.

Second, monitor whether these systems cause more false positives, such as delayed withdrawals or extra KYC checks, which are a trade-off between convenience and security.

Third, keep assuming scammers will adapt. Even with better AI on exchanges, many scams try to keep victims on fake apps or wallets so stolen funds never enter a regulated venue that can block them.

Conclusion

Bybits claim that its AI systems blocked around $300 million in attempted scams illustrates how security is shifting from purely rule-based checks to adaptive, data-driven defenses. For crypto users, the main takeaway is that venue choice and personal security both matter: robust exchange AI can dramatically reduce losses when scammers try to cash out, but the safest outcomes still start with avoiding the scam in the first place.