Thus we are not going to talk about the consequences that the new wave in AI might have for the empiricism/rationalism debate (see Buckner 2018), nor are we going to consider—much—the question of whether it is reasonable to say that what these programs do is ‘learning’ in anything like the sense with which we are familiar (Buckner 2019, 4.2), and we’ll pass over interesting questions about what we can learn about philosophy of mind from deep learning (López- Rubio 2018). We are not going to talk about the clearly very important ethical issues involved, either the recondite ones, science- fictional ones (such as the paperclip maximizer and Roko’s Basilisk (see e.g. Bostrom 2014 for some of these issues)), or the more down- to- earth issues about, for example, self- driving cars (Nyholm and Smids 2016, Lin et al. 2017), or racist and sexist bias in AI resulting from racist and sexist data sets (Zou and Schiebinger 2018). We also won’t consider political consequences and implications for policy making (Floridi et al. 2018).