Agent Failures

Put aside complex legal considerations for the moment and ask two simple questions:

Why would users be dissatisfied?

1. the user doesn't know what they want and are ambiguous in their preference expression,
2. if the user knows what they want but the preferences are not captured well, or
3. if the agent operated upon well expressed preferences with poor results. 

What are the possible modes of failure of the agent? Was the computer agreement based on:

1. direct human involvement?
2. a preconfigured rule set by the human?
3. a preconfigured rule set by a trusted authority?

Should computer agreements include information on how involved the human actually was in making the agreement?