AML POLICY
Accordingly, new AML software, products
and tools with AI technology are quickly
entering the market to assist FIs with com-
plex AML activities, such as identifying and
evaluating suspicious activity. For instance,
existing AI technology has the ability to
track the behavior of individuals and enti-
ties involved in potential money laundering
activity and link them to one another
through collective intelligence gathering
and machine learning techniques that
apply adaptive learning rather than prede-
fined scenarios and rules. 18 This method
can provide a stronger and more compre-
hensive understanding of behavior over
time. Other machine learning-based plat-
forms have the ability to identify and aggre-
gate voluminous and unstructured data
points, such as narratives within SARs or
publicly available information, to assist
with better recognizing AML risks and
automating KYC processes. 19
Taking a leap: Go bold to get ahead
The evolving partnerships with technology
companies, such as IBM and the many vendor
products entering the market, illustrate the
applications and benefits of cognitive tech-
nology in the AML space, particularly as tech-
nology becomes cheaper and the demand for
skilled labor increases. The value of cut-
ting-edge technology and “digital labor”
combined with the growing importance of
collaboration and centralization presents
opportunities for how FIs can begin to imple-
ment, experiment with, or simply think about
how to leverage cognitive computing capabil-
ities through a transformative framework or
methodology that encourages communities
that are against financial crime to operate as
partners in a common cause.
For instance, a centralized cognitive com-
puting tool that incorporates multiple FIs (in
a structured, yet cooperative fashion similar
agency). By identifying an appropriate
third-party candidate with sufficient tech-
nological capabilities and dedicated
resources, the organization would be able
to provide oversight, as well as set up and
pilot the tool in a safer and more focused
environment. This form of centralization
may also be more cost-effective, time-effi-
cient and supportive for FIs, as it allows the
effort to be shared and distributed. For
instance, this unit could maintain gover-
nance over the tool and assist with coordi-
nating the retrieval of information from
participants, as well as cleansing and ano-
nymizing data where necessary.
to crowdsourcing, where participants collec-
tively contribute to, and benefit from, pooled
information and knowledge) fits well with
the ongoing emphasis on collaboration, cen-
tralization, and investing in technology. A
basic benefit would be a more powerful data
set (e.g., aggregated KYC and transactional
activity) with an integrated feedback loop
that links back to the originating FIs, allow-
ing for enhanced data quality and analysis.
The tool could serve as a supplemental data
store that FIs can use side by side with
existing mechanisms to validate or comple-
ment current data; and/or provide future
potential to feed directly into the FIs’ sys-
tems and/or models.
While similar tools exist in the form of
vendor products and AML collaboration
software, an option of taking this even fur-
ther is a semi-communal cognitive comput-
ing tool that is managed centrally by an
independent third party (such as a self-reg-
ulatory organization or a government
Althou