Criminals mimic voices of loved ones using AI: how to protect yourself from financial fraud

The development of artificial intelligence technologies creates new threats for private individuals. Criminals have learned to generate convincing audio recordings and videos that mimic the voices and faces of real people. This allows them to manipulate victims into transferring funds or disclosing personal data by impersonating relatives, friends, or company representatives.

The mechanism of fraudulent schemes

Modern AI tools make it possible to create forged content that mimics the behaviour and voices of real people. A victim may be convinced they are communicating with a known individual because the voice sounds familiar and the behaviour meets expectations. Fraudsters exploit this trust to extract financial resources, passwords, or other confidential information.

Five key signs of potential deception

There is a range of specific behavioural patterns that often accompany such attacks. Although none of these factors individually constitutes definitive proof of fraud, their combination should serve as a signal to exercise heightened caution.

• Pressure and demand for urgency. The caller attempts to create a sense of panic, leaving no time to verify information.
• Demand for secrecy. The fraudster convinces the victim not to inform other family members or acquaintances about the situation.
• Insistence on transferring funds. This often involves urgent payments or the purchase of gift cards.
• Request for confidential data. This may include bank details, passwords, or one-time confirmation codes.
• Unfamiliar phone number. Even if the number appears familiar, this does not guarantee the authenticity of the call.

Identity verification: proven methods

If you receive a call from someone claiming to be a bank employee or representative of another institution but using a new number, do not rely solely on the stated position or the persuasiveness of the voice. The safest strategy is to end the conversation and independently call the organisation using the official number, which can be found on the back of a bank card or on the company's website.

In cases where an alleged relative or friend asks to urgently transfer money, it is necessary to contact them via an alternative method: using a number saved in contacts or through a different messenger. It is also advisable to clarify the situation with other close relatives.

Family code word as protection against deepfakes

An effective protective tool is a code word or control question agreed upon in advance with family members. The answer to it should be known only to the closest people and must not contain information available in open sources.

If the caller cannot answer the control question correctly, any financial requests should be ignored until the person's identity is fully verified. It is worth remembering that a voice, video image, and phone number are not reliable identifiers in themselves.

Actions in case of encountering fraud

In the event of suspected manipulation, it is necessary to immediately cease communication and refuse to provide any data. If funds have already been transferred, the bank should be notified urgently to block the transaction or protect the account. It is also important to contact the police, preserving all evidence: phone numbers, correspondence, call recordings, and payment details.

Additionally, the person whose voice or name was used by the criminals must be warned. This will allow them to promptly alert other relatives about the threat.

The main principle of security is to avoid haste. Even if the situation appears urgent and the voice sounds familiar, one should pause, verify the identity of the caller, and only then make a decision.