Living with AI, Research Horizons (2018)
Powerful AI needs to be reliably aligned with human values. Does this mean that AI will eventually have to police those values? Huw Price and Karina Vold consider the trade-off between safety and autonomy in the era of superintelligence.