Bitget Wallet flags 37,000 malicious signing requests weekly in Q3, up from 5,000

Bitget Wallet flagged an average of about 37,000 potentially malicious signing requests a week between July and September, 7.4 times the 5,000 weekly average in the second quarter.

The self-custodial wallet provider said the jump may reflect more attacks, better detection, or both, and that it cannot say how much each factor contributed. It announced on Wednesday that it is expanding AI-powered threat detection.

AI detection takes a bigger role in identifying malicious websites

The expansion comes as cryptocurrency scammers increasingly rely on convincing websites, fraudulent applications and deceptive transaction requests to gain access to users’ digital assets.

Rather than depending exclusively on previously identified malicious websites and external threat intelligence databases, Bitget Wallet is expanding its use of AI to examine suspicious activity more closely.

Its enhanced detection system evaluates website code and domain registration information to identify characteristics associated with fraudulent platforms. These checks can help identify suspicious websites that imitate legitimate cryptocurrency services.

The wallet also uses transaction simulation to examine what a signing request would authorize before a user approves it. This approach is particularly relevant because some malicious transactions conceal permissions that could allow attackers to transfer tokens or gain control over users’ assets.

The company’s October figures indicate that artificial intelligence is becoming a more significant part of its security operations.

Bitget Wallet reported that AI identified approximately 90% of newly blocked risky domains during October, compared with 36% in the second quarter.

Another notable finding was that nearly 70% of the domains identified by its AI systems during October had no corresponding alerts in the external threat intelligence feeds it checked.

This suggests that proprietary detection systems may help identify suspicious websites before they appear on established security databases. However, the figures are based on Bitget Wallet’s internal monitoring and do not independently establish the overall effectiveness of AI-based security across the cryptocurrency industry.

The latest Crypto phishing alerts also demonstrate why wallet providers are increasingly investing in preventive security measures rather than relying solely on warnings issued after suspicious activity has been discovered.

Recent crypto scams highlight the risks of signing malicious transactions

Bitget Wallet’s latest security update arrives amid several incidents demonstrating how easily cryptocurrency holders can become targets of phishing operations.

In July 2026, a cryptocurrency trader lost approximately $1 million after approving a malicious token transaction on the Ethereum network. According to reporting by Cointelegraph, the victim lost 999,999 USDT after signing a fraudulent token approval.

The incident illustrated how attackers can exploit transaction permissions to drain funds from a wallet. In approval-based scams, victims may believe they are interacting with a legitimate service while unknowingly granting malicious smart contracts permission to access their tokens.

A separate investigation published by cybersecurity company Rapid7 in August uncovered another extensive operation targeting cryptocurrency users.

The investigation identified approximately 885,000 phone numbers stored within infrastructure connected to a phishing campaign. The operation used deceptive communications and websites impersonating established cryptocurrency wallet providers to target potential victims.

Rapid7’s findings also revealed how attackers used account-validation tools to identify cryptocurrency users, making their targeting efforts more precise.

These incidents demonstrate that phishing operations are evolving beyond traditional fraudulent emails. Attackers are combining impersonation websites, phone-based social engineering and malicious transaction approvals to exploit weaknesses in user awareness and wallet security.

Consequently, Crypto phishing alerts are becoming an important part of the broader effort to protect digital asset holders.

AI security could become essential for crypto wallet providers

The latest developments suggest that cryptocurrency wallet security is moving towards more proactive threat detection, particularly as malicious actors develop increasingly sophisticated methods of targeting users.

For self-custodial wallets, the challenge is especially important because users retain control over their private keys and must personally authorize transactions. Although this model provides greater control over digital assets, it also places considerable responsibility on individuals to understand the permissions they grant.

AI-assisted monitoring could help reduce this exposure by identifying suspicious websites, examining transaction behaviour and warning users about potentially dangerous approvals before funds are compromised.

However, automated detection should not be treated as a complete solution. Users still need to verify website addresses, avoid unfamiliar links, examine transaction permissions and remain cautious when interacting with unsolicited messages.

Bitget Wallet’s latest figures provide an indication of how much its detection capabilities have expanded, but the distinction between suspicious interactions and confirmed attacks remains important when evaluating the scale of the threat.

As cryptocurrency adoption continues to expand, Crypto phishing alerts could become an increasingly important measure of how wallet providers identify and respond to emerging risks.

For now, Bitget Wallet’s decision to expand AI-assisted monitoring reflects a wider industry priority: detecting fraudulent activity before users sign transactions that could put their digital assets at risk.