Criminals can perfectly mimic a loved one’s voice to trick you and steal your money using deepfake technology, the National Cyber ​​Security Directorate (DNSC) warned on Monday, issuing a guide on how to identify and protect against the phenomenon. This year, a pensioner from Vaslui lost 52,000 lei, and in 2019, with the help of a Deepfake audio file, the CEO of a British company was defrauded of 220,000 euros.

• VIEW THE DEEPFAKE GUIDE

The danger of Deepfake: the technology can be used to commit fraud or manipulate public opinion

Deepfake technology uses advanced artificial intelligence techniques to create a fake image or audio-video content that can be used in fraud attempts, such as impersonating you during phone or video calls to gain unauthorized access to information or financial resources.

• “One example: there are several cases where a public figure is on video giving financial advice. In this video, a person strongly recommends investing in a certain company or business, promising high returns and minimal risks.
• Details matter: The video is professionally made with high quality picture and sound. The public persona comes across as sincere and persuasive, using accessible language and persuasive arguments.
• Campaign Goal: Such a Deepfake can be used to manipulate the public into investing in an illegal business (scam1) owned or controlled by the manipulators. Impact: Videos go viral on social media with huge potential to be viewed by millions of people. Many of those who see the video may be persuaded to invest in the recommended business at the risk of losing significant amounts of money later on.” warns DNSC.

Deepfake technology can be used to create false records that implicate people in activities they did not perform, or to obtain false evidence in a legal context.

It can also be used in fraud attempts, such as spoofing your identity during phone or video calls to gain unauthorized access to information or financial resources.

In politics and other fields, Deepfakes can be used to manipulate public opinion, discredit opponents, or create confusion. They can influence elections, international relations and fuel conspiracy theories.

Real life examples of using Deepfake: fake content with Isărescu and people who lost money

DNSC reminds that the National Bank of Romania (BNR) recently alerted the public about a fraud scheme involving the use of Deepfake technology to create fake videos of the Governor of the BNR.

In these videos, the governor allegedly advertises an investment platform, but the NBR recognized them as fake. The scam uses artificial intelligence to change the governor’s voice and image to mislead the public into participating in fraudulent investments with the promise of quick and easy financial gains.

Another case is a pensioner from Vaslui, who became a victim of a scam posted on the Internet on the YouTube platform. Fraudsters have created a fake video in which a famous banker and other famous personalities recommend an investment platform.

The promise of a quick profit convinced the pensioner to invest 52,000 lei, savings accumulated over 20 years of work. Even if the woman reported this fraud to the authorities, experts believe that the chances of recovering the lost funds are minimal.

“Criminals are becoming more resourceful. Now I can perfectly imitate the voice of a loved one to fool you. You may get a call pretending to be a family member who urgently needs money. We recommend that you always verify the identity of the caller, even if it is a known relative or friend. Never send money in a hurry and notify the authorities immediately if you have any suspicions. Only through vigilance and information can you protect yourself from this form of fraud,” DNSC warns.

In 2019, Deepfake audio was used to defraud a CEO of €220,000.

• “The managing director of a British energy company thought he was on the phone with the managing director of its German parent company when he executed an order to immediately transfer €220,000 to the Hungarian supplier’s bank account. The voice actually belonged to a fraudster who used artificial intelligence voice technology to impersonate the German CEO,” DNSC said.

The danger of Deepfake in the electoral context

Deepfake technologies capable of realistically synthesizing images and voices can significantly influence the opinion and votes of voters during election campaigns.

• “A politician’s reputation can be seriously damaged by fabricated Deepfake videos, which can discredit his image and damage his chances of winning the election. The spread of misinformation via Deepfake can manipulate public perception of a politician’s character and agenda, negatively impacting his or her political career.
• Deepfake can be used as a strategic tool to discredit competing parties by creating false material that portrays them in a negative light,” the DNSC manual says.

How to recognize a dipfake: speech/sound and lip movements are not synchronized when pronouncing the letters b, m and p

DNSC experts say there are certain signs that can detect a Deepfake, here’s what to look out for:

• Environment (e.g. no shadows, too many reflections, blurry areas)
• Facial imperfections (unrealistic moles, unsynchronized blinking, distortions inside the mouth such as missing teeth and tongue, too perfect teeth, etc.)
• Out of sync with speech/sound and lip movement eg: due to sneezing.

“When pronouncing the letters b, m and p, you can notice a lack of synchronization of speech/sound and lip movement. Sometimes gray pixels appear on the edges of changed components. Distinguish whether it is a fake and when the person in the recording is viewed from a different angle. If photos of a person from different angles were not used to create the Deepfake content, the algorithm would not be able to determine the person’s appearance from a different angle, leading to distortions.” says DNSC.

Tips to avoid being scammed by Deepfakes

Avoiding Deepfake scams requires a combination of healthy skepticism, attention to detail, and the use of verification tools. Here are some helpful tips:

• Don’t believe everything you see on the Internet! The Internet is a huge source of information, but not all of it is true. It is important to develop a healthy skepticism and carefully analyze any video or photo content before accepting it as genuine.
• Look for signs of manipulation: Deepfakes can be very sophisticated, but they can often be recognized by certain signs. Look out for lighting discrepancies, misalignment, uneven skin, or lip-sync issues.
• Check the source: Where is the video or image from? Is it distributed on a reliable platform? Seek confirmation of information from reliable sources or directly from the organizations or individuals involved.
• Use the verification tools: There are many organizations and online tools that can help you verify that information is true. Use them to verify the authenticity of suspicious content.
• Don’t rely on just one source: Seek confirmation from multiple reliable sources. A single video or image is not enough to confirm a piece of information.
• Learn about Deepfakes: The better you understand how this technology works, the better you can spot fakes. There are many online resources that explain the principles of Deepfakes and detection methods.

By applying these tips, you can reduce the risk of being scammed by Deepfake content and promote a culture of verification and responsibility online.

What to do if you are a victim of a Deepfake

If you find yourself in the unfortunate position of being a victim of a Deepfake, it’s important to act quickly and effectively to minimize the damage. Here are some steps you can take:

• Document the abuse: Keep copies of Deepfake content, including URLs, screenshots, or any other form of evidence that may be appropriate. This is important for any legal action or subsequent reporting.
• Content of the report: Most social media platforms and websites have strict policies regarding Deepfakes and manipulated content. Users can easily report suspicious content using the platform’s built-in reporting feature.
• Contact the authorities: In serious cases where Deepfake content violates laws regarding defamation, harassment, or distribution of pornography without consent, you may need to contact local authorities or other law enforcement agencies.
• Request for legal assistance: Consult an attorney to evaluate your legal options. This may include legal action against those who created or distributed the Deepfake content.
• Use online reputation management services: There are companies that specialize in improving your online presence and removing or reducing the impact of negative content. These services can be useful in protecting your image in the long run.
• Communicate with transparency: If Deepfake could affect your career or personal relationships, consider talking openly about the situation with your employer, colleagues, or loved ones. Providing context and your side of the story can help reduce the negative impact.
• Protect your personal information: After the Deepfake incident, it’s important to be extra vigilant about online security. Check your social media privacy settings, change your passwords, and monitor your account activity for signs of unauthorized access.
• Emotional support: The psychological trauma suffered by a Deepfake victim can be significant. Don’t hesitate to reach out to friends, family, or mental health professionals for support.
• Education and awareness: Help raise awareness about the dangers of Deepfakes by sharing your experience if you feel comfortable. This can help inform and protect others.

Read more: Isarescu reported to Zuckerberg about “Isarescu” for deep fake / Meta response: Community standards are followed