AI Safety in a World of Vulnerable Machine Learning Systems

Link post