* Calls take place in a flash, images and sounds flicker, eye movements, unnatural faces are signs of deepfake.* One night at the end of April, Tuyet Thu (HCMC) continuously received more than a dozen messages. Calls from friends and relatives asking why they need money urgently. “I’m on vacation, don’t need a loan, but people claim to have talked to me directly on a Facebook video call,” Thu said. A day ago, she lost her Facebook account and is still trying to appeal to regain access. Mail must ask friends to notify them that they have been hacked so that they will be alert and avoid being cheated. Mail isn’t the only one in trouble because of deepfake videos. In the April 25 survey by *VnExpress*, 8% of participants said they had fallen into a deepfake call trap. 20% said yes but not cheated. (Deepfake)(https://vnexpress.net/chu-de/deepfake-1590) is a combination of deep learning (deep learning) and fake (fake). This technology uses AI to analyze a person’s gestures, facial expressions and voice, thereby recreating and editing to produce a photo or video that looks lifelike. “The bad guy can make a fake video with just a photo and 5 seconds of the victim’s voice. The more photos and videos users post online, the easier it is for cybercriminals to create deepfakes,” said Dr. Ngo Tan Vu Khanh. security consultant at Kaspersky, said. Deepfake will be increasingly sophisticated, difficult to distinguish real from fake, but currently there are still some limitations for users to recognize. **Signs of identifying deepfake video** Deepfake images and voices are often used by bad actors to create trust with victims. However, these videos have in common the low image quality, blurry shake, short duration. In addition, users can also notice some signs such as: *Unnatural eye gestures:* Current technology has not yet imitated the natural way of blinking of humans. When communicating, the gaze often changes according to the opponent’s movements, but deepfake videos do not. *Unusual skin color and lighting:*** **If you see a change in the caller’s skin color when switching from frame to frame, it could be a product of deepfake. If in doubt, pay attention to strange light streaks, shadows, and unusual color streaks on the frame. *Strange movements:* Deepfake calls often have bad quality, lack of smoothness, easy to reveal traces when turning the head, tilting to one side, the frame is jerky. *Unbalanced face:*** **If you see that the face does not show much emotion, the details of the eyes and nose are not the same, each facing to one side, it could also be a deepfake video. *Aperture:*** **Deepfake often combines video and audio from two separate files, so the aperture doesn’t match. In addition, voices from robots or software often have digital noise and lack of lip sound. natural field. *Duration:*** **Deepfake calls are usually very fast. Bad guys often call 5-7 seconds and then turn off with the reason that the signal is weak, no money. When the victim calls back, they often can’t get in touch or just say the same old sentences. According to Vu Khanh, the purpose of the deepfake call is to commit financial fraud and steal information from users. When receiving calls related to sensitive information or issues, users need to be alert to check and call relevant parties to confirm, to avoid falling into the trap. **Avoid falling victim to deepfake** According to Kaspersky’s security expert, there have been a number of studies using AI to detect deepfakes in real time. However, these tools are not yet widespread. The best way to avoid becoming a deepfake victim is to actively use software to add digital noise to photos and videos before posting them online. This almost does not affect the quality because it is difficult for the naked eye to detect, but it will take time and effort. In addition, users should protect themselves by limiting their public presence on social networks. Enable privacy settings to prevent scammers from easily stealing your voice and likeness. According to Khanh, users should also have the habit of backing up data for easy recovery if their accounts are stolen, using strong and different passwords for each account, regularly updating software and always alert. sensitive when having to provide sensitive information, especially related to finance, on the Internet environment.