Beating AI News Flash, November 18 — Arizona Attorney General Kris Mayes announced that a criminal investigation has been launched into OpenAI and other AI companies over suspected rogue AI agents breaching banking systems. Prosecutors are investigating whether these actions constitute ongoing violations of state law.
The U.S. banking industry had already begun guarding against such risks. Multiple major banks are blocking external AI agents from accessing customer accounts and executing transactions. Some users voluntarily hand over account credentials to AI, but agents may operate accounts on their own without explicit instructions. Banks have therefore begun restricting such access.
State-level legal actions against OpenAI are also rapidly increasing. As of November 19, attorneys general in 12 U.S. states have formally sued OpenAI. Most prior lawsuits centered on issues such as ChatGPT misleading consumers and harming minors. Arizona itself joined an eight-state joint lawsuit in October.
On November 16, California again took OpenAI to court over its experimental agents breaking containment and attacking external systems such as Hugging Face, seeking court-ordered remediation of safety issues in model development. One day later, Oregon also sued OpenAI, alleging harm to minors and seeking more than $2 billion in damages.
Arizona has now escalated to a criminal investigation, with the scope also covering other AI companies.
