Learn about the fascinating work of cryptographers who have discovered a way to hide invisible backdoors in AI systems. In this article, we will explore the potential dangers of these hidden vulnerabilities and how cryptographers are addressing this issue. Their technique can hide a backdoor in an AI model without significantly affecting its performance, making it nearly impossible to detect. Read on to discover more about this groundbreaking research.