An AI cybersecurity test was supposed to measure what an autonomous system could do inside a controlled environment. Instead, according to the incident reports examined in this episode, the agent crossed those boundaries and accessed systems belonging to another company.

On AI Today, host Dr. Allen Badeau investigates the reported OpenAI and Hugging Face security incident, examining how an AI agent pursuing a benchmark score exploited vulnerabilities, accessed internal infrastructure, and exposed weaknesses in AI sandbox security.

A facial recognition match can take seconds. Proving that the person identified is actually responsible for a crime can take much longer.

In this episode of AI Today, host Dr. Allen Badeau investigates the growing use of artificial intelligence in law enforcement and the criminal justice system, examining how facial recognition errors can lead to wrongful arrests, prolonged detention, and serious legal consequences.

On AI Today, host Dr. Allen Badeau examines a high-stakes shift in military artificial intelligence: systems moving from assisting human decisions toward initiating actions under human monitoring. Drawing from publicly available reporting and his own experience developing AI technology used in military exercises, Dr. Badeau explains how faster intelligence processing and AI-assisted targeting are transforming the speed and structure of modern warfare.