Every data scientist working with deep learning needs to consider the consequences of their work on society and think about human issues such as how to avoid runaway feedback loops, ensuring an appeals process for humans impacted by their algorithms, and adequately explaining the constraints of their algorithms to people using them.