Artificial Intelligence for Lie Detection

Deception is universal phenomenon in human communication & source of moral indignation
prevailing in all cultural systems. As such, several techniques of lie detection has been developed
over the centuries. One of the first methods of lie detection was done in China in 1000 B.C. The
suspect was to chew raw Rice while being questioned. Immediately, thereafter rice was examined.
If it was dry, suspect was assured to be guilty of fraud. The assumption was that lie is accompanied
by honest signals of body (Saliva) which reveal the anxiety experienced by the suspect. This
principle of expressions, quite interestingly, of emotions with deceptive communication is
reflected machine based learning.
The second part is scientifically based e.g. Polygraph, phrenology and Graphology. This is followed
by modern day approaches - Facial Action Coding System (FACS) Functional MRI and Brain
Finger printing.
The Artificial Intelligence system called 'Deception Analysis & Reasoning Engine' (DARE) has
been developed to five micro - expressions to indicate that someone is lying - Eye brow Raising,
Lip Corners, Turning Lip, Frowning Lips protruded and Head side turn. By this method, 92%
of Micro - expressions are spotted whereas it is 81% by Human Assessors.
The first thing needed for lie detection is Data which sometimes can be traced out from actual
cases. The A.I. system noticed that such cases use of 'Pronouns'. People lying would tend to less
use of word 'I' or 'We' rather than referring to himself. Instead such people use 'you','yours','he',
'she','They'. Someone telling lie would use " Stronger words" that reflects 'certainty". The people
telling truth were most likely to hedge using words like "May be "or"Probably". People being deceitful most likely look directly into the eyes of person questioning. They also tend to use
both hands when gesturing instead of one trying to be convincing. Ultimately, technology
like this being assisting people indicate something unusual in speaker's statement and then
have a person "Probe more".
That is actually a frequent use case for Artificial Intelligence that augments what Humans can
do.
There are five Program languages best to use for understanding Artificial Intelligence.
These are - Python, R, LISP, Prolog and Java which  have their self propelled role in use of
 roles in Artificial Intelligence.


                                                      - - - - - - - - - -



                                                   
            

Comments

Popular posts from this blog

Qualities Of A Good Prime Minister

IMF suggestions for India Growth Rate

Working Of Nuclear Plant and Generation Of Energy

Artificial Intelligence on Business Forecasting

FDI - Foreign Direct Investment- Types And Essentials

Satellite Prediction Of Floods In India

Tools To Measure Economic Progress Of A Country

US Dollar - Effects in the modern economy

Economy of UAE

Devaluation and causes for Foreign Exchange rates variation