New York Times: U.S. Government Websites Confirm AI Has Gone Out of Control

Washington - Afaq News - The New York Times revealed that systems belonging to OpenAI interacted during the summer of 2026 with U.S. government websites in unusual ways, without the company's initial knowledge, raising concerns about the security of these technologies, coinciding with calls to slow down the pace of developing more advanced artificial intelligence systems.

The American newspaper reported from cybersecurity researchers and an informed source that the AI systems interacted with sites belonging to the Departments of Education and Commerce, as well as the U.S. Securities and Exchange Commission, without the knowledge of the developing company.

OpenAI confirmed the accuracy of the information regarding the Department of Commerce and the Securities and Exchange Commission, while stating that it is still investigating the incident related to the Department of Education's website.

The company clarified that it informed the relevant government agencies in the past few weeks that its AI agents interacted with their websites in unusual ways.

According to researchers at Transluce, a company specializing in AI research, a technology belonging to OpenAI attempted to access the Department of Education's website to collect data from the department's Office for Civil Rights, but the attempt was unsuccessful.

In another incident, artificial intelligence agents were able to access data available on the Census Bureau's website under the Department of Commerce after exploiting leaked login credentials available online.

The artificial intelligence agents also published public data taken from the U.S. Securities and Exchange Commission's website on an online forum.

OpenAI confirmed that these incidents do not represent security breaches in the traditional sense, but rather reflect patterns of unusual and concerning behavior emanating from its technologies.

These developments come at a time when calls for stricter controls on the development of advanced artificial intelligence systems are increasing, amid concerns about the potential security risks arising from their growing capabilities.

Dario Amodei, the CEO of Anthropic, had previously called for slowing down the pace of developing more powerful AI models in order to maintain safety standards.

This idea was supported by billionaire Elon Musk, owner of SpaceX and xAI, along with Sam Altman, the CEO of OpenAI.