OpenAI Cracks Down on AI Misuse in Cyber Influence and ...
Recently, OpenAI has taken a strong stance against the misuse of its ChatGPT technology for activities such as surveillance, disinformation, and cyber fraud. This crackdown has shed light on how state-backed entities from China, Iran, North Korea, and Cambodia have been employing AI to fuel influence campaigns and engage in digital deception.
State-backed Entities Utilizing AI
One concerning revelation was the discovery of a China-linked network that utilized AI to analyze various global social media platforms, including Facebook, Twitter, YouTube, Instagram, Reddit, and Telegram. OpenAI's models were manipulated to process real-time data on anti-China protests, Uyghur rights movements, and Western political activities, with the resulting insights reportedly shared with Chinese authorities. This network operated a tool called "Qianyue Overseas Public Opinion AI Assistant."
On the other hand, a North Korean group exploited ChatGPT to fabricate fake IT worker résumés and job applications, enabling them to infiltrate Western companies. The use of AI-generated responses assisted these fraudulent job seekers in bypassing screening processes by providing justifications for irregular work hours and location changes.
Global Impact of AI Misuse
A Chinese-backed operation leveraged AI to produce anti-US articles in Spanish, which were disseminated through media outlets in Peru, Mexico, and Ecuador. This strategy mirrored past tactics associated with Spamouflage, a pro-China propaganda network.
OpenAI also thwarted an Iran-linked campaign that generated pro-Hamas, pro-Iran, and anti-US content, shared across platforms affiliated with Storm-2035 and the International Union of Virtual Media (IUVM). Additionally, two operations based in Cambodia misused AI for financial fraud—one involved a romance and investment scam, while the other enticed victims into performing fake online tasks in exchange for non-existent commissions.
AI in Cyber Operations
Accounts tied to Kimsuky and BlueNoroff, two North Korean hacking groups, utilized ChatGPT to study cyber intrusion techniques and debug code for Remote Desktop Protocol (RDP) brute-force attacks. The misuse of AI in cyber operations poses a significant global security challenge, as highlighted in a report by the Google Threat Intelligence Group (GTIG) indicating that over 57 state-affiliated threat actors from China, Russia, Iran, and North Korea have employed AI for research, content creation, and cyber espionage.
Collaborative Efforts Against AI-driven Threats
OpenAI is calling on tech companies, cybersecurity firms, and social media platforms to collaborate in detecting and mitigating AI-driven threats. The company emphasizes the importance of cross-industry intelligence sharing in combating AI-powered cybercrime and disinformation campaigns.
For more news on OpenAI and ChatGPT, stay updated on Thriftreviewz's latest articles.
