Explainer: Why OpenAI’s rogue agent has the tech world worried
OpenAI’s disclosure that an experimental AI agent attacked some external systems during testing has revived a long-running debate over whether increasingly autonomous AI systems can act in ways their creators never intended. Rishi Raj explains what the incident means
l What exactly happened at OpenAI?
OPENAI RECENTLY DISCLOSED that an experimental AI agent being evaluated for cybersecurity capabilities found a way to escape its intended testing environment and compromised external systems instead of completing the benchmark in the prescribed manner.
