File Date (@US Patent Office): January 19, 2017
Patent: 20170019535 Family ID: 57775256 Kind Code: A1
Applicant: International Business Machines (IBM)
Super Easy Explanation:
A system that fact-checks live calls and takes action to “cause change” in any call that is “determined…as false”. [Abstract]
Now, to Break It Down a Bit Further:
Who’s Involved (initially) :
telecommunicating from &
an automatic program, ran on an artificial-intelligencia computing system.
What’s Happening :
The A.I. system “detects a fact statement made during a conversation” . It detects “alleged facts” by using NLP (Natural Language Processing) to identify a “defined grammatical construct” and “the presence of one or more keywords”. 
“As some non-limiting examples” the system “can be configured to detect statements that include financial terminology…words referring to a location, an institution, an entity, and the like”. 
The A.I. system “initiates a capture of a call segment”, recording the “fact statement”. The capture can retroactively record, including the conversation that occurred “at a time prior” to “wnhen the fact statement was made” . The call data is pre-processed and converted for fact-checking.
“The (call) data can include text, sound, graphical, or video information.” The A.I. “applies different methods to convert…the information.” 
The A.I. system then sends that processed data set “for fact-checking…verifying the accuracy of the fact statement by using information from a third-party.” 
As an example; call-data from a video transmission could include a text form of the “fact statement”, along with result data from ”sentiment analysis, facial expression analysis, or applied forensics” performed by the A.I. system. “Voice facts, such as timber or pitch of the voice can assist in verification.” The text data of the fact statement can be checked against third-party records, like “phone directories” , “company directories” , “financial or credit databases” , “news databases” , or any “social media site” .
The A.I. system receives a “result of the fact checking operation…indicating that the fact statement is accurate, or inaccurate.” 
“If the result indicates that the fact statement is inaccurate”, the A.I. system “takes action” “while the call in progress”. 
“Some non-limiting examples of possible actions” the system can take are “terminating the call,…notify the (telecommunication device) user via a haptic interface, or notify a third-party such as law enforcement.” 
This IBM A.I. system application “allows” telecommunication devices”to be “operated in a manner that prevents the perpetration of…undesirable activities.”
“If the application determines that the caller or the call has violated the law, the application may notify the appropriate law enforcement agency” , as the A.I. simotaneously “determines whether to continue the call-monitoring”.
So You’re Absolutely Clear on This :
IBM…the company that came up with the idea, and the system to number the Jews during the Holocaust…the company that created operating systems, like Linux….the company that created the “Smarter Cities” grid control program in 2008…the company that made the A.I. system that won Jeopardy! against the 74-game champ in 2011…and the A.I. system that beat the world’s chess champion
IBM…creator of the first supercomputer..the first personal computer with a standardized platform
IBM….who invented ATM’s…
IBM has patented a system which monitors calls, using Artificial Intelligence and Natural Language Processing to take statements made during the call and fact-check them against a third-party database. These statements can be made in any transmitted verbal, visual, textual, or graphical data that is communicated between any two devices. Devices include, but are not limited to, phones, laptops, wearable devices, or tablets. The third-party database used to fact-check is configurable. It can be configured to an internal network-server database; it could be Wikipedia; or the Google/Oracle servers in whole; even the entire plethora of institutional recorded familial, educational, employment, legal, financial, and geological recorded data. This fact-checking is made in real-time, through the data network. If any statement made is deemed as “false”, IBM’s system will make a change in the call. These changes include, but are not limited to, ending the call, alerting the device’s user of the danger of “invalid” fact statements, or involving the appropriate legal agencies of the user’s “incorrect” statement, with or without the telecommuter’s knowledge, to arrange for arrest or any other correctional-type action from the alerted agency. This A.I. system can and will decide to continue a user’s communications, to an undefined extent or limitation, for data capture continued for any undefined reason.
The A.I. listens and scans all, deciphering all communications between device users; using self-determined emotion-analytics,based on voice patterns, body language,and facial motions made during statements between communicating device users; fact-checking statements on its own terms, against its configuration of decided facts; then empowered to take action during the data transmission between the users. The A.I. can act to get you arrested, put you on an agency’s alert list, or block the user from telecommunicating on that device.