Self-learning AI enhances behavioral analytics for detecting advanced threats and insider risks, but demands careful attention to bias, transparency, and ethical use.

Scientists at Massachusetts Institute of Technology have devised a way for large language models to keep learning on the fly—a step toward building AI that continually improves itself.