Increasingly easy access to tools for easy creation of deepfake images, videos and audio is bringing huge negative impacts on companies as well. The Analytics Data Factory, a Czech company that focuses on real-time fraud prevention, estimates that a large proportion of Czech companies will fall victim to a deepfake attack by the end of next year. The consequences cannot be predicted, as it will depend mainly on the ability of employees to recognize a deepfake in time.
“We have always had the advantage of the complexity of the Czech language when it comes to detecting cyber fraud. It was often enough to focus on grammatical errors and nonsensical phrases in fraudulent emails that tried to lure sensitive information or access data. But the use of artificial intelligence tools is gradually eliminating this advantage. At the same time, deepfake applications are becoming increasingly accessible and easier to use even for laypeople. Their use will grow significantly and by next year at the latest, every company will encounter some form of deepfake-based cyber fraud,” says Jiří Mojžíš, fraud management expert and CTO of Analytics Data Factory.
The most common type of deepfake attack targeting companies will be forging a supervisor’s voice on the phone. Attackers can easily find the necessary contacts, for example, on a website and call from a spoofed phone number, where the artificially created yet familiar voice of the supervisor can give commands, for example, to transfer funds or pass certain information out of the company. Similarly, deepfake videos can be used in online meetings.
Fake corporate statements will also be a major threat when a deepfake video appears on social media in which a manager or employee accuses competitors, business partners, or “reveals” false internal company information, for example. Even if the authenticity of the video is immediately refuted, it can spread for years across the internet and significantly damage a company’s reputation and operations. Similarly, deepfake nudity – i.e., artificially created videos or photographs – can be used to attack a company’s employees directly on a personal level.
“Just as deepfake tools are constantly improving, so are the tools for detecting them. Still, the biggest role here will be played by employees – educating them about deepfake should be the goal of every company. It is employees who will increasingly encounter deepfake fraud. And if, for example, an accountant enters a payment order based on a combination of a forged email and a phone call from a company director, it’s usually too late to do anything about it. Prevention is important,” says Lukáš Benzl, director and founder of the Czech Association of Artificial Intelligence.
Deepfake is a term for realistic image or sound editing. It uses advanced computer processing by artificial intelligence to modify, for example, a person’s face, facial expressions and speech. People then perform actions in the video that they don’t actually do and say things they never said.
Due to the ever-increasing threat of deepfake, Analytics Data Factory worked closely with the Czech Association of Artificial Intelligence to create the DEEPFAKE 2024: Defensive Strategy for Czech Companies manual, which is freely available for download on the association’s website below:
