OCBC Bank fights money laundering with AI double act
Singapore’s OCBC Bank has piloted two fintech solutions from a pair of start-ups as it looks to combat money laundering and terrorism financing via artificial intelligence (AI).
The two firms involved – BlackSwan Technologies and Silent Eight – were part of OCBC’s second fintech accelerator programme, Open Vault, which was launched last year.
Both companies used AI to conduct research on individuals and entities suspected of illegal financing. This is done by searching for information on individual profiles, and mapping how suspicious transactions may be linked to one another to see if they are fraudulent or illegal.
Alex Ng, head of OCBC Bank’s group transaction surveillance that led the pilot, says Open Vault gave it the “opportunity to work closely with the fintechs during a short burst of three months, to experiment by providing them with anonymised real data to see how the solutions can meet our needs” – and the bank was “not disappointed” in the results.
OCBC says at present the AML process is carried out manually by an internal compliance analyst, with the total time required lasting between one hour and a few days.
Suspicious transaction monitoring is usually done in two parts. Namely, issues are flagged to a compliance team; who then research by scanning social media accounts, newspapers, search engines and databases to create a dossier of the individual, to get an understanding of banking patterns, associates and behaviours. The bank says this process can take up to one hour or more – and then followed up with mapping to find possible patterns of connections.
Silent Eight helps to automate the desktop research process by enabling analysts to digitally scan internet search engines, news sites and internal and external databases to put together the dossier within a minute. AI enables the system to filter out irrelevant information such as individuals with identical names.
BlackSwan Technologies uses AI to analyse the suspicious transactions, mapping them to a network of related transactions to identify possible connections with other individuals or companies which may be atypical or previously unknown. This can simplify the task of the analyst and provide leads to follow up on.