“AI Clone & Deepfake” – Beware of 2026 New Scam Tactics as Voices and Faces of Your Family May Not Be Real

As we enter 2026, Generative AI technology is expected to advance further compared to previous years. However, technologies for positive purposes may be misused by criminals, turning them into a double-edged sword. Throughout 2025, scammers didn’t stop at traditional phone-call fraud asking victims to transfer money; they began using AI voice cloningto imitate voices and deceive victims instead. Moreover, in November 2025, authorities arrested a group of Chinese scammers who used AI to even trick AI systems, bypassing banks’ security measures known as biometric verification. Although this method has not yet been used to scam people in Thailand, it is believed that such technologies will certainly be used in 2026.

What was scammers’ AI like in 2025 ?

Pol. Col. Neti Wongkularb, Deputy Commander of the Technology Crime Suppression Division under the Central Investigation Bureau (CIB), told Thai PBS Verify that one AI-involved scam tactic used by criminals in 2025 was AI-based voice impersonation. This trick has already been used for several months. Scammers often begin with the pretense that they have changed their phone number, asking victims to delete the old number and save the new one that the criminals are calling from. The scammers then engage in casual conversation, usually hanging up first and waiting 1–2 days before making a second call. During this second contact, they begin creating a story to deceive the victim.

[caption id="" align="aligncenter" width="1200"]

Pol. Col. Neti Wongkularb, Deputy Commander of the Technology Crime Suppression Division, the Central Investigation Bureau (CIB)[/caption]

What stories did criminals use to deceive victims in 2025 ?

Pol. Col. Neti explained that in 2025, criminals used a variety of stories to scam victims, but all of them began the same way. The victim is first told to delete the old phone number of the person the scammer is impersonating, with the excuse that they have changed to a new one. After that, the scammers wait about one to two days before contacting the victim again. Based on observed behavior, they usually do not attempt to scam the victim immediately on the first strike. Instead, they allow some time to pass, then call back with a story involving urgent trouble, especially financial problems—such as:

• Claiming they have made a purchase but cannot contact their spouse, asking the victim to transfer money to the shop, promising to repay immediately once they are able to reach out to their spouse or relatives.
• Claiming to be the victim’s child or parent, contacting the victim to ask for a money transfer.

In both cases, victims are deceived into transferring ten thousand to hundred thousand of baht, or even more if the scammers are successful.

How to observe if it's someone you know or a scammer ?

To determine whether an incoming call is from a scammer, consider the following checks:

• If it’s someone you know and have already saved in your contacts, even if the caller’s voice sounds exactly like your friend, parent, sibling, child, or grandchild, be suspicious if the call comes from an unfamiliar or strange number and the caller claims they have a new number and asks you to delete the old one. This is a scam tactic.
• Politely hang up the call from the unfamiliar number immediately, even if the caller claims to be your child, relative, or someone you know. Then, call back using the original number you already have saved and ask whether they called you. These two steps are among the simplest and most effective ways to protect yourself from scammers.

An example of a criminal impersonating an acquaintance.

Online threats in 2026

The next level of scams: from deep-fake voices to breaking identity verification systems

In the past, we may have seen AI deep-fake voices used to impersonate acquaintances and ask for loans. Today, however, there are attempts to use high-resolution AI-generated faces that can move in response to bank identity-verification prompts—such as blinking, turning left or right, or opening the mouth. These actions are key security checkpoints when transferring large sums of money.

A recent example was the arrest of a group of Chinese scammers in November 2025, who used AI to deceive AI to bypass banks’ biometric security systems.

[caption id="attachment_7974" align="aligncenter" width="1024"]

Chinese scam operations use AI to deceive AI and bypass bank security systems[/caption]

Although this arrest did not involve the creation of AI deepfakes to scam Thai people, it is believed that such technology carries a high risk of being used to deceive Thai victims in 2026. Thai PBS Verify found that these tools have been developed to be more sophisticated. Scammers typically gather their “raw materials” from:

• Front-facing facial photos that individuals have posted on social media
• Videos or short clips showing faces from multiple angles
• Personal data from leaked databases, which can be used to forge applications or reset passwords

Example: AI clips used to fool AI

Why do scams involving “relatives or acquaintances” continue to work so effectively ?

Independent criminologist Worapat Sirikiratayanon shared his perspective on why impersonation of relatives or acquaintances continues to be an effective scam tactic. He explained that today many people no longer work in their original hometowns. Some move to major provincial cities, others come to work in Bangkok, and some even work abroad.

Due to economic pressures and living conditions that force people to work in different locations, families that once lived together as extended households now experience gaps and separation. These gaps create vulnerabilities that criminals use as key opportunities to commit crimes—by impersonating family members or close friends who were once trusted and using those relationships as the primary tool to deceive victims.

[caption id="attachment_7970" align="aligncenter" width="1024"]

Mr. Worapat Phuengpong, independent scholar in criminology and victimology[/caption]

“I still believe that this type of scam remains effective to a certain extent, and that criminals will continue to rely on it as a core method for reaching new victims. However, in the future it may become difficult to tell the difference. This can be seen in the way scammers are using deepfakes to fake faces, voices, or even video calls. I believe this approach will continue, but with technologies or software that make it appear even more realistic.”

This is evident from developments reported in 2025, when there was news about software capable of generating images with much greater realism. Such technologies may be increasingly used for fraud. That said, voice impersonation without showing a face—where victims cannot see pixel-level abnormalities—will likely continue as well. This method remains highly effective, because it requires low investment and allows scammers to target large numbers of victims in a short period of time without using complex technology. It is then likely to be one of the most widely used scam methods in 2026.

Three key signs of “AI identity” you need to watch out for

Even though AI can appear very realistic, careful observation during a video call or identity verification may reveal the following abnormalities:

• Unnatural movement around facial edges:Look closely at the hairline or the boundary between the face and the neck. You may notice blurring or glitches when the person moves quickly.
• Eye depth and light reflections:Reflections in AI-generated eyes often do not match the actual lighting direction in the environment, or the blinking may appear stiff and unnatural.
• Audio–visual sync issues:The voice may be delayed compared to lip movements, or the tone may sound monotone, lacking natural human emotion.

Protect yourself: close the door on identity theft

The Cyber Crime Investigation Bureau (CCIB) and cybersecurity experts recommend the following safety practices:

• Set a family-only passcode:If you receive a suspicious video call or a request for money from someone claiming to be a close relative, ask for a “secret code” or a detail that only the two of you would know to confirm their identity.
• Avoid posting high-resolution, front-facing photos:Sharing overly clear selfies in public spaces effectively hands AI high-quality “source material” that can be easily processed and misused.
• Be alert to suspicious data requests:Banks do not have a policy of sending links asking customers to upload facial videos or biometric data through any application other than the bank’s official app.
• Use multi-factor authentication (MFA):In addition to facial recognition, enable other verification methods, such as an authenticator app, for added security.

Facial identity verification technology used by Thai banks has now incorporated a ‘Liveness Detection’ system to prevent fraud using photos or videos. However, scammers are also developing AI technologies to counter these systems.

Therefore, in 2026, the most important defenses remain ‘staying calm’ and ‘thinking carefully’ which continue to be the strongest safeguards, especially when facing scammers who are evolving their methods of deception.