Page 26 - GSTL_16 April 2020_Vol 35_Part 3

J46 GST LAW TIMES [ Vol. 35
manual interface of the officer with the assessee which needs to be avoided in
order to have a transparent, corruption free system.
Many tax authorities in countries like Australia, Canada, Norway and
United Kingdom have started to build AI-supported models to predict the high-
risk individuals/entities that are most likely to react positively to different tax
authority interventions for recovering revenue. These techniques, which draw on
behavioural sciences as well as data mining are being pursued as a priority. AI
has opened up opportunities for new, simple policy options which were not pos-
sible earlier. For example, the Australian Taxation Office (ATO) way back 2002
created and implemented an integrated information technology platform, consol-
idating the ATO’s core functions into two systems - namely Integrated Core Pro-
cessing (ICP) which is used for client registration, processing and accounting and
Siebel which is used for case management, collections and customer relationship
management. Apart from updating internal systems, the ATO has integrated
Nuance’s intelligent virtual Assistant, Nina, within the organization’s website in
2016. This allows ATO’s customers to have their questions answered by AI,
called Alex. Having Alex handling the ‘easier’ questions allows agents to focus
on more complex cases that delight customers, reduce costs and deliver signifi-
cant benefits. Artificial Intelligence in auditing can be used to identify potential
tax fraud cases. Referred to as “predictive modeling,” machine learning applica-
tions are now being used by tax agencies to identify cases having characteristics
that could indicate potential fraud. It often helps find subtle clues hidden in
mounds of data that are sometimes missed or overlooked by auditors.
When the appropriate data showing non-filing of returns or short pay-
ment of duty on a supply classified under a particular chapter heading, or inap-
propriate claim of ITC or excess/fraudulent refund say on account of non-
existent initials of the assessee, or such like things are noticed, the automation
(AI) may be a preferred choice in coming days. With the widening of assessee
base it would be appropriate to nudge or send the notifications indicating the
error/mistake committed by the assessee using the algorithm and creating an AI
based programming. This is in consonance with the intention of the Government
also which does not prefer to penalize the taxpayer if the dues are paid within a
stipulated time.
Thus, a fair combination of behavioural science, Data Mining and AI is
the need of the hour. As per the news reports a meeting of the seniormost officers
of C.B.I. & C. and CBDT to augment the use of technology including AI was re-
cently (January, 2020) convened by the head of the Revenue Department is a wel-
come step in this direction.

_______

GST LAW TIMES  16th April 2020  146