The rise of deepfake AI has raised great concerns among internet users. This technology has been utilized to commit various cybercrimes recently, often without the perpetrators being traced. A recent notable cybercrime involving deepfake AI occurred on the centralized exchange OKX.

According to Cointelegraph, an account holding approximately $11 million on OKX was drained by hackers using deepfake AI technology. This account belonged to the pseudonymous operations manager at Quant Matter, Crypto Lala. Surprisingly, Crypto Lala did not receive any email or phone notifications for new device access, making it appear as though the exchange itself was responsible for the theft.

Star Xu, the founder of OKX, revealed that an advanced "coin-stealing hacking gang" used deepfake AI to bypass facial recognition security systems. This allowed the hackers to gain direct access to accounts. Xu also highlighted the audacity of these criminals, who misled victims into believing the exchange was at fault for the theft.

Deepfake AI can bypass facial recognition in centralized exchanges through several steps:

1. Data Acquisition: Hackers collect images or video footage of the target from social media, public databases, or other sources. The more data they have, the more accurate the deepfake will be.

2. Deepfake Creation: Using deep learning algorithms and neural networks, hackers create a highly realistic deepfake video or image of the target. This involves training a model to replicate the target’s facial movements, expressions, and other distinguishing features.

3. Authentication Attempt: With the deepfake ready, hackers use it to impersonate the target during the facial recognition process. For example, during a login attempt on a centralized exchange, the deepfake video or image is presented in place of the real person’s face.

4. Bypassing Security: If the facial recognition system is not robust enough to detect deepfakes, it accepts the fake as genuine and grants access to the account. This breach allows hackers to control the account, transfer funds, and carry out other unauthorized activities.

To counter these threats, exchanges need to implement multi-factor authentication, use advanced deepfake detection algorithms, and continuously update their security protocols to recognize and prevent sophisticated attacks. Additionally, users of centralized exchanges should be cautious about how they expose their faces on social media platforms.

Is It Still Advisable to Use Centralized Exchanges to Store Funds?

Personally, I have always advised against using centralized exchanges to store funds. There is nothing safer than having your funds in a place where you have full control. While getting a Ledger might be costly, it is worth it. Using a self-custodial wallet is also desirable if you cannot afford a Ledger. Centralized exchanges should be used solely for trading, fiat-to-crypto transactions, and vice versa.

As cybercrimes involving deepfake AI technology increase, internet users need to be cautious about their online exposure. Centralized exchanges must enhance their security measures to detect deepfakes effectively..