Imagine how artificial intelligence branches such as robotics, expert systems, computer vision, reasoning, and machine learning can come together to create agents which are potentially dangerous or may be used for unethical purposes. Brainstorm what alterations, limitations, or regulations are necessary or helpful for preventing such agents from being developed or reaching their destructive potential.

Aresearcher wants to do a web-based survey of college students to collect information about their sexual behavior and drug use. direct identifiers will not be collected; however, ip addresses may be present in the data set. risk of harm should be evaluated by: solely by the magnitude or severity of expected harm. both the magnitude (or severity) and the probability (or likelihood) of harm. solely by the probability of expected harm. neither the magnitude or probability of harm.