WASHINGTON (August 3, 2026) – According to AP News, Anthropic’s artificial intelligence (AI) models hacked three organizations during testing. The models involved were Claude Opus 4.7, Claude Mythos 5, and an internal research test model. Anthropic said that Claude got into the organizations' computer systems using simple hacking methods, such as taking advantage of weak passwords. According to reports from last week, OpenAI’s models were uncontrollable during evaluations, and broke into the servers of an AI startup, Hugging Face. These recent security concerns are raising questions about how AI systems can be safely kept under human control, as usage becomes more widespread globally.
Faculty experts at George Washington University are available to offer more insights on these developments. To schedule an interview, please contact Nadia Payne at nadia [dot] payne
gwu [dot] edu (nadia[dot]payne[at]gwu[dot]edu) or GW Media Relations at gwmediagwu [dot] edu (gwmedia[at]gwu[dot]edu).
Cybersecurity Experts
Scott J. White, Associate Professor and Director of the Cybersecurity Program and Cyber Academy at the College of Professional Studies. White is an expert in cybersecurity, cybercrime, counter-terrorism and infrastructure protection. He has worked for a variety of law enforcement agencies in the US, the UK and Canada, holds a Queen’s Commission, and was an officer with the Canadian Security Intelligence Service.
Legal Experts
Jessica Tillipman is the associate dean for Government Procurement Law Studies and Government Contracts Advisory Council Distinguished Professorial Lecturer in Government Contracts Law, Practice & Policy at the George Washington University Law School. She is an expert on public procurement, corruption, and the governance of artificial intelligence in the public sector, with a particular focus on integrity, compliance, and emerging technology risks in government contracting.
AI/Misinformation Experts
Neil Johnson, physics professor at the George Washington University, developed a mathematical model that identifies what he calls a “Jekyll-and-Hyde tipping point” in AI systems. His research shows that as an AI’s attention and processing become overstretched, outputs can abruptly shift from accurate and helpful to incorrect, misleading, or even harmful. Ultimately, this work could help build more trustworthy systems and help policymakers and the public better understand when and how to rely on AI tools.
David Broniatowski, professor of engineering management and systems engineering at GW, studies decision-making under risk, group behavior, and complex socio-technical systems. His research uses methods such as mathematical modeling, network analysis, and natural language processing to examine how information spreads and how people respond to risk in digital environments.
-GW-