OpenAI, the creator of the AI chatbot ChatGPT, has teamed up with its primary investor, Microsoft, to thwart five cyberattacks orchestrated by various malicious entities.
Microsoft has been monitoring hacking groups affiliated with Russian military intelligence, Iran's Revolutionary Guards, and the governments of China and North Korea. According to a report released on Wednesday, these groups have been exploring the utilization of large language models (LLMs), powered by artificial intelligence, in their hacking endeavors.
OpenAI revealed that the five cyberattacks were attributed to two Chinese-linked groups: Charcoal Typhoon and Salmon Typhoon. Additionally, attacks were traced back to Iran via the Crimson Dust Storm, to North Korea through the Emerald Sleet, and to Russia via the Forest Blizzard.
The hacking groups attempted to employ ChatGPT-4 for various malicious activities, including reconnaissance on companies and cybersecurity tools, debugging code, scripting, phishing campaigns, translating technical documents, evading malware detection, and researching satellite communications and radar technology. Upon discovery, the accounts associated with these activities were promptly deactivated.
While OpenAI successfully averted these incidents, it acknowledged the ongoing challenge of preventing malicious exploitation of its AI programs. Despite efforts to implement safeguards, hackers have found ways to circumvent these measures and manipulate ChatGPT to generate harmful or inappropriate content.
In response to the escalating concerns surrounding AI security, OpenAI and other leading entities have joined forces with the Biden administration to establish the Artificial Intelligence Security Institute and the American Artificial Intelligence Security Institute Consortium (AISIC). These initiatives, mandated by President Joe Biden's executive order on AI security issued in late October 2023, aim to foster safe AI development, combat AI-generated deepfakes, and address cybersecurity risks.
