Amer Sharaf, CEO of the Cybersecurity Systems and Services Division at the Dubai Electronic Security Center (DESC), warned about the threats posed by deepfakes and new fraud methods. In his speech at the global GISEC event in Dubai, he emphasized that artificial intelligence technology enables criminals to significantly improve their fraud methods.
Changes in Fraud Methods
Sharaf stated that criminals are now able to use artificial intelligence to identify information about their potential victims and generate convincing messages, calls, and videos. He also noted that new threats such as phishing, voice fraud, and deepfakes have become increasingly sophisticated due to the use of artificial intelligence.
He emphasized that to combat these threats, individuals should double-check: "Every email you receive in your work should be re-verified through your usual methods before making any transactions. Contact the individual and ensure the information is correct. Especially regarding fund transfers, go to the person's office."
Using Technology for Cyber Defense
Sharaf also referred to new technologies that help improve cyber defenses. Notably, DESC unveiled an artificial intelligence model named Saraab on the first day of GISEC, designed to identify deepfake videos. This tool analyzes each frame of video and uses a heat map to detect signs of manipulation.
Saraab will be made available to the public as an open-source tool, allowing individuals to download and use it. This technology has been introduced as a useful tool to the community, with hopes that it can play an effective role in identifying cyber threats.
In this context, other experts at the GISEC event also pointed to the dangers arising from the information individuals input into public AI services. Jansen De Hert, a recognized instructor at the SANS Institute, stated that users often do not understand how their information is processed in AI tools. He said, "The number one mistake is that people use it publicly without understanding the type of AI they are using."
He noted that data entered into public services may be used to train AI models, and this is one of the biggest immediate dangers. For this reason, there is a need for regulation and management of artificial intelligence.
