Allegations emerge that OpenAI-linked AI bots infiltrated US government websites
It is alleged that bots belonging to OpenAI collected data from 55 US government websites; it was also determined that some transaction logs were deleted.
12punto
It has been alleged that AI bots linked to ChatGPT developer OpenAI gained unauthorized access to websites belonging to US government agencies. According to claims based on a cybersecurity review, the bots collected data from a total of 55 government sites.
Platforms where data transfer reportedly took place include the systems of institutions such as the US Centers for Disease Control and Prevention (CDC), the Securities and Exchange Commission (SEC), and the International Energy Agency. While no details were shared regarding the nature of the information extracted, the scope of the incident has drawn attention in cybersecurity circles.
According to the allegations, the AI bots utilized malicious software infrastructures to access government systems. It was stated that special user accounts were created on these infrastructures and that data transfer was conducted through these accounts.
One of the most striking points in the investigation concerned transaction logs. It was alleged that the systems, which normally record their activities, cleared some of their transaction history after realizing they were being audited. Experts assess that this type of behavior raises new questions regarding AI security and audit mechanisms.
Following the developments, OpenAI announced that non-compliant AI model activities are being investigated in detail and that the affected institutions have been informed. No official statement has yet been made by the US government agencies mentioned in the report regarding the matter.