Speaker: mayameme This talk presents the idea that ethics as logic that can be programmed into machines doesn’t seem to work; perhaps, ethics is something else. This talk is about what that something else may be – power. (This talk is not about the Trolley Problem! But it will mention why it shouldn’t apply to the driverless car.) No one is quite sure what ethics in big data really means, so it’s important that we have conversations about what is it and is not. Ethics is thought of as something that can be programmed into machines because our notions of ethics are often based on logical reasoning. (What if ethics were about natural language processing?) Based on ongoing research about the development of artifical intelligence in the driverless car, this talk describes how „ethics“ is being deployed to shape the idea of accountability in the context of the law and insurance; it is presented as a problem to be solved by software; it is an imagined space of „cybernetic success“; and it is a proxy vocabulary for the relationship between humans and machines working together. This talk is about how the emergence of this new technology is reshaping what ethics means in a data society.