The CSA’s Singapore Cyber Landscape 2023 report, released on Tuesday, highlights that malicious actors are leveraging generative AI for deepfake scams, bypassing biometric authentication, and identifying vulnerabilities in software. Deepfakes, which use AI techniques to alter or manipulate visual and audio content, have been employed for commercial and political purposes. This year, several Members of Parliament received extortion letters featuring manipulated images, and Senior Minister Lee Hsien Loong warned about deepfake videos misrepresenting his statements on international relations.
Traditional AI typically performs specific tasks based on predefined data, analyzing and predicting outcomes but not creating new content. This technology can generate new images, videos, and audio, exemplified by ChatGPT, OpenAI’s chatbot.
AI has also enabled malicious actors to scale up their operations. The CSA and its partners analyzed phishing emails from 2023, finding that about 13 percent contained AI-generated content, which was grammatically superior and more logically structured. These AI-generated emails aimed to reduce logical gaps and enhance legitimacy by adapting to various tones to exploit a wide range of emotions in victims.
Additionally, AI has been used to scrape personal identification information from social media profiles and websites, increasing the speed and scale of cyberattacks.
The CSA cautioned that malicious actors could misuse legitimate research on generative AI’s negative applications, incorporating these findings into their attacks. The use of generative AI adds a new dimension to cyber threats, making it crucial for individuals and organizations to learn how to detect and respond to such threats. Techniques for identifying deepfakes include evaluating the message, analyzing audio-visual elements, and using authentication tools.
Despite the growing sophistication of cyberattacks, Singapore saw a 52 percent decline in phishing attempts in 2023 compared to the previous year, contrary to the global trend of rising phishing incidents. However, the number of phishing attempts in 2023 remained 30 percent higher than in 2021. Phishing continues to pose a significant threat, with cybercriminals making their attempts appear more legitimate. In 2023, over a third of phishing attempts used the credible-looking domain “.com” instead of “.xyz,” and more than half of the phishing URLs employed the secure “HTTPS protocol,” a significant increase from 9 percent in 2022.
The banking and financial services, government, and technology sectors were the most targeted industries in phishing attempts, with 63 percent of the spoofed organizations belonging to the banking and financial services sector. This industry is frequently targeted because it holds sensitive and valuable information, such as personal details and login credentials, which are highly attractive to cybercriminals.
