As we aim to deploy “AI for All” in sectors such as healthcare and education, it is essential that these systems are not only secure but also safe and responsible. A narrow focus on security alone is insufficient