10 REGULATING, NOT UNPLUGGING
We cannot realistically go back to a world without AI, so we need to find ways to live safely with it. Researchers are working on a list of principles that every agent will have to respect. This will need to be enforceable and verifiable, so the most fundamental property that we should expect of an agent is “auditability”: being designed in a way that enables some form of inspection. It will then be possible to expect safety, fairness, privacy, transparency and other important requirements that legal scholars are debating at this time, so that we can trust intelligent machines. The shortcut that gave us the present form of AI makes this step difficult, but not impossible: replacing ...
Get The Shortcut now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
