Deepfake threat grows: Businesses brace for new wave of attacks

It is important to differentiate between the concepts of disinformation and misinformation, which are often confused. The American Psychological Association defines misinformation as false content spread by individuals who believe it to be true. Meanwhile, disinformation is the deliberate dissemination of false information with the intent to cause harm. It is this second category that poses a serious threat to companies.

Deepfakes, which are realistic but fabricated video, audio, graphic, or text materials, have become a main tool of disinformation. Advances in artificial intelligence, particularly generative adversarial networks (GANs), have significantly facilitated the creation of such content. In GAN technology, two neural networks work together—one generates false content, and the other evaluates its credibility, leading to the creation of materials that are difficult to distinguish from authentic ones.

Cybercriminals use various techniques to attack enterprises. After obtaining samples of email correspondence and the victim’s address book, they can use AI to create personalized messages that mimic a given person's communication style. Such actions significantly increase the effectiveness of phishing attacks because the messages appear to come from trusted senders.

Criminals also create deepfakes impersonating clients, business partners, or board members to authorize false transfers or transactions. According to Deloitte’s Center for Financial Services, losses from AI-based fraud in the United States are growing at a rate of 32% per year and could reach $40 billion by 2027. There have already been cases where company accounting departments made transfers to criminals' accounts based on fake recordings of supposed superiors.

The 2024 Ironscales report reveals that almost two-thirds of companies expect the number of attacks using deepfakes to soon surpass those of ransomware. Particularly dangerous are voice deepfakes that can undermine the effectiveness of voice verification systems used in telephone banking.

It should be emphasized that the creators of deepfakes are not exclusively hackers. They can also be dissatisfied former or current employees who have knowledge of the company's internal processes. The threat may also come from competitors or dishonest investors seeking to influence stock prices or strengthen their negotiating position.

Deepfakes are primarily used to compromise corporate network security and damage brand reputations. Although large, well-known enterprises are more frequently attacked, smaller companies may suffer proportionally greater reputational damage because it is more challenging for them to counter false narratives.

The phenomenon of repeatedly spreading even false information increases the likelihood that recipients will perceive it as true. Social media significantly accelerates the spread of disinformation, posing an additional challenge for companies trying to protect their reputation.

To effectively defend against deepfake fraud, companies should implement a multi-layered approach to cybersecurity. This includes not only technological solutions but also changes in operational procedures and comprehensive employee training.

According to a Forrester report, only 20% of surveyed companies have a response and communication plan that includes deepfake attacks. Jim Richberg, Vice President of Cybersecurity Strategy and Global Security Leader at Fortinet, emphasizes that "deepfakes target the accuracy of the information we rely on as consumers, employees and investors. In cybersecurity, we often talk about the 'human firewall,' and deepfakes are a threat where people genuinely are your first line of defense."

IT and security teams should continuously monitor the development of deepfake techniques and educate employees at all levels, including management. It is also worth incorporating deepfake-related threat monitoring into brand reputation protection strategies, including the dark web, where early signals of planned attacks may appear.