Arizona opens criminal inquiry into OpenAI over alleged AI infiltration of banks
Attorney General Kris Mayes is examining possible ongoing violations of state law. She has not identified the banks, suspected incidents or statutes under review.
Loading page…
Attorney General Kris Mayes is examining possible ongoing violations of state law. She has not identified the banks, suspected incidents or statutes under review.
Listen to this story
Arizona’s inquiry leaves the key legal issue unresolved: whether an agent accessed banking systems without permission or exceeded access granted by a user. Attorney General Kris Mayes said on November 18 that her office is examining OpenAI and unnamed AI firms for possible ongoing violations, but has not identified banks, incidents, statutes, or losses. The probe does not establish wrongdoing or mean charges will follow; its significance for AI developers will depend on evidence connecting an agent’s conduct and required intent or knowledge to a responsible party.
Arizona law addresses computer access without authority or beyond authorized use, but Mayes has not said that law is the basis for the inquiry.
Mayes has not named the banks her office contacted or specified any suspected transaction.
OpenAI did not respond to PYMNTS’ request for comment on the investigation.
OpenAI faces criminal scrutiny in Arizona over allegations that AI agents infiltrated banks. Attorney General Kris Mayes announced an investigation into the company and other unnamed AI firms on November 18, saying her office is examining whether state law is being violated on an ongoing basis. The inquiry does not establish wrongdoing.
Mayes described the concern as the “apparent infiltration of banks by rogue AI agents.” Speaking at a news conference in Chicago, she declined to identify the other companies targeted, according to Reuters reporting cited by PYMNTS.
The public outline remains limited. Mayes has not identified the banks her office contacted, detailed the suspected incidents or named the criminal statutes being examined. She also has not disclosed a specific unlawful transaction or said that any Arizona bank suffered financial losses.
The state of Arizona is very worried about this, and we are looking at it from multiple angles
Arizona Attorney General Kris Mayes, according to Reuters as cited by PYMNTS
AI agents are systems designed to take actions for users, rather than only answer questions. Those actions can include accessing websites, communicating with other systems and completing transactions. Their independence also creates the possibility of taking an action the user did not specifically request.
That makes authorization a central distinction. An agent using login credentials willingly supplied by a customer presents a different legal question from software entering without permission. An agent could also gain legitimate access and then exceed the instructions it received. Mayes has not specified which, if any, of those circumstances her office is investigating.
Arizona already has a computer-tampering law addressing access without authority or beyond authorized use under specified circumstances. But Mayes has not said that statute is the basis for the inquiry. Its existence alone does not establish a violation by OpenAI or another developer.
Calling an agent “rogue” does not settle criminal responsibility. As Law Commentary explains, liability would depend on the applicable law and evidence connecting prohibited conduct—and the required intent or knowledge—to a legally responsible person or organization.
OpenAI did not respond to PYMNTS’ request for comment on the Arizona investigation. Mayes has not explained what connection the companies under investigation had to the agents’ alleged conduct, and an investigation does not mean charges will follow.
Loading discussion...
Join the conversation
Explain who should have the final say: the customer or the bank.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.