Still another, expected to be popular among both politically savvy consumers, news outlets, policy wonks, lobbyists and others, examines congressional voting behavior.
“The computer would take a particular bill and look to see if there is anything strange or different in the voting,” Hammond said. “If you have somebody you wouldn’t expect vote for it, then the question is, what are the reasons? The system goes and looks. We are building a system aimed at being able to create news stories around every single vote that takes place, and can explain why that vote has taken place in that particular way.”
The overreaching goal is for computers to go beyond the straightforward process of collecting and synthesizing information to a level of greater sophistication that enables them to “get” the subtleties within the context of the data, much as humans try to do.
“We want to humanize the machine to get it working for you, to give the machines intelligence, which they really don’t have right now,” Hammond said.“We want the machine to understand when you are asking a real question, a rhetorical question or even making a joke—the machine will get it.”