CAREER GUIDANCE
are many smart tech-oriented professionals creating these tools , we will start seeing the AML professional ’ s job change and it is hard to predict exactly what that change will be . Chris Phillips , senior vice president and director of AML at Valley Bank , uses a great analogy to describe how AI will catalyze the evolution of the AML professional . He says , “ By way of example , a farmer in late-1800s America could not conceive of earning a living sitting behind a desk in a skyscraper ; they simply had no way to conceive of such a position . There were a series of social and technological changes between that farmer and the businessperson of the 1950s that needed to occur before the transition could happen . We are in a similar position , in that we will not know what the next generation of jobs will look like until we get there , until technology matures and companies start to rethink what they are willing to pay employees to do . Between here and there , people need to be curious , constantly looking to learn new skills and understand how technology is changing our world .” 3 Although it will be difficult , let us continue to take a stab at what future FCC programs will look like from a personnel standpoint .
How AI will affect AML jobs and programs
AI is expected to become increasingly sophisticated in the broader fight against financial crime . This evolution will manifest in highly complex forms , but more importantly for FCC programs , it will eliminate the more administrative and tedious responsibilities that AML professionals ― at all levels ― still deal with now . What follows is a discussion of how this transformation will unfold .
Enhanced pattern recognition and data sorting
AI systems will leverage deep learning techniques to identify complex patterns in transactional data that are indicative of money laundering activities . This capability will help in flagging suspicious activities that human analysts might miss . According to Caruso , “ The first impact of AI will be on all AML / compliance jobs whose primary function is gathering , sorting , organizing and presenting data . Many are copying , pasting and preparing to conduct basic due diligence , sanctions or adverse media alert reviews . These jobs will not exist in 10 years , likely much sooner . If work is rote , repetitive and follows easily definable steps , don ’ t count on these jobs existing .” 4
Quicker responses and alerts
Although it will take time ― and the hard work of data scientists and governance , risk and compliance professionals ― AI will enable continuous , real-time monitoring of financial transactions . This will allow FIs to detect and respond to suspicious activities as they happen . In the meantime , however , AI will still produce alerts faster than most programs currently do . In addition , when AI detects potential financial crimes , it will generate and escalate alerts , complete with detailed analysis and suggested actions . In Phillips ’ s opinion , “ In the very near future , much of the work done now by investigators will all be done automatically . However , it will be presented to the investigator for review and adjudication . Once an activity has been adjudicated , the platform will autogenerate the narrative for sign-off and filing . The investigator will still be necessary , as they will be needed to recognize new patterns , as well as to adjudicate intent on the part of the suspect , but we will need fewer of them .” 5
False positives reduction
False positives are the bane of many BSA officers and AML programs . They are costly to remediate from both a technological and personnel standpoint . AI will refine its algorithms to reduce the number of false positives . AI systems will continually learn from feedback , adjusting their models to become more accurate over time , further reducing false positives . Phillips adds , “ The reduction in false positives will come from changes in how alerts are generated . Rather than using static deviation percentages or dollar amounts , the detection will be based on anomaly detection with better trained and tuned datasets .” 6
This will lead to the need for not only more data scientists , but also more QA professionals . Stephen Sargeant , founder of Airdropd , a media production house that works with Web3 and traditional finance in everything related to blockchain and compliance , says : “ You will see many more people in charge of conducting [ QA ] on the outputs from AI . You will also see new departments strictly in charge of training the AI and LLMs . In the next 10 years I still think AI will be used as a supplement . I think every employee will have an AI ( robot / machine ) that they will use throughout the day to be their ‘ thought partner .’ They will be using this machine to bounce ideas off and to qualify thoughts and decisions and ― more importantly ― AI will be used to create consistencies in processes and procedures . Nobody has time to look back through 400-page manuals to decide on how they are going to report a sanctioned transaction . They will now use an AI machine to walk them through the process , so no matter who trained you within the organization , the procedure will be easy to follow .” 7
68 acamstoday . org