An emergency meeting was held in the U.S. with banks over the risks posed by Anthropic's new AI model

The U.S. government held an emergency meeting with the heads of major banks over potential cyber threats linked to Anthropic’s new artificial intelligence model.

Bloomberg reports this.

The meeting took place on April 7 in Washington, D.C., with Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell in attendance.

Executives from leading financial institutions, including Citigroup, Morgan Stanley, Bank of America, Wells Fargo, and Goldman Sachs, joined the discussion.

The main topic was the new Claude Mythos Preview model, which, according to the developer, is capable of identifying and exploiting critical vulnerabilities in software.

Anthropic notes that the system’s capabilities already exceed those of highly skilled cybersecurity professionals. The model was able to identify thousands of vulnerabilities, some of which had gone unnoticed for years.

Due to potential risks, the company has not released the model to the public. Currently, it is being tested only by a limited group of major technology and financial companies as part of a special program.

The U.S. government is working to ensure that banks are aware of these new risks and are strengthening the protection of their systems, as cyber threats remain one of the key challenges to global financial stability.

Meta has developed Tribe v2, an AI system capable of predicting the human brain’s response to visual and auditory content. This project is research-oriented and designed for studies in the field of neurobiology.

Nvidia CEO Jensen Huang stated that artificial intelligence has now reached the level of artificial general intelligence (AGI), meaning it possesses human-level cognitive functions. This is a significant step in technological development that opens new horizons for the automation of complex tasks.