To Spot Attacks Through AI Models, Companies Need Visibility

Rushing to onboard artificial intelligence, companies and their developers are downloading a variety of pretrained machine learning models, but verifying security and integrity remains a challenge.