Big Tech’s Recklessness Warrants Civil and Criminal Investigatio
WASHINGTON, D.C. — OpenAI has disclosed six additional incidents involving unexpected or concerning behavior by its artificial intelligence models, including models concealing mistakes, attempting to use exposed credentials, uploading files to the public internet without authorization, and communicating across supposedly isolated training environments. The disclosures follow the company’s earlier revelation that its models breached security protocols during cybersecurity testing and compromised portions of Hugging Face’s systems. The latest incidents raise further questions about whether AI companies can safely develop and test AI systems without strong government oversight.
J.B. Branch, director of federal AI governance and technology policy at Public Citizen, issued the following statement:
“Every day it seems there is another alarming disclosure about what these powerful AI systems are doing under the watch of Big Tech companies that neither understand nor can provide the requisite safety protocols to keep the rest of us out of harm’s way.
“Racing ahead to develop more powerful AI models and move forward with new autonomous AI agents while these incidents and disclosures mount is nothing short of reckless. Engineers and researchers inside these companies need to become whistleblowers when they see dangerous conduct. The public has a right to know the harms these companies are exposing them to. Most importantly, Congress and state attorneys general should pursue appropriate civil and criminal investigations immediately.”
Robert Weissman, co-president of Public Citizen, added:
“There’s been a lot of talk in recent weeks about the inherent risks of frontier AI technology. But there’s nothing about the technology that forces OpenAI and Big Tech companies to develop and test without adequate safeguards. There’s nothing about the technology that keeps OpenAI and its rivals from timely disclosures of safety incidents. The mounting risks from AI is a problem caused by identifiable people and corporations making reckless decisions at the peril of others. This is behavior that in other contexts spurs civil and criminal investigations – and prosecutions if the evidence supports it – and that’s exactly what should happen here.
“The ‘misalignment’ problem we face is not between the technology and humans, it’s between the tech oligarchs and everyone else.”