Looking At Artificial Intelligence With 2021 Vision
A new year is upon us, and artificial intelligence is climbing to the top of the world’s most impactful technologies. In fact the growth of AI worldwide increased 154% last year.1 Moreover, revenues of software that leverage AI climbed to over $9 billion.2 Consequently, as we’ll discuss below, firms that adopt AI are gaining a competitive advantage, which enables them to dominate their respective industries. It’s no wonder why this phenomenon has captivated everyone’s attention—AI is making groundbreaking developments.
AI Under the Microscope
In the medical world, serious technological advancements are being developed by analyzing data to provide personalized support for managing chronic illnesses. Cigna Corp., a global health service company, recently announced plans to release an artificial-intelligence-based system that will identify disparities in the treatment of chronic diseases. This new method of intervention has the potential to save lives by using algorithms to evaluate clinical data and generate personalized recommendations. For Cigna the integration of AI has, undoubtedly, bolstered competitive advantages by decreasing inaccuracies and reaching higher degrees of precision.
Similar to the complexities in the medical world, facets of the due diligence process pose a threat to producing accurate background checks. For example, background checks conducted on individuals with family names, such as John Adam Thompson III,3 pose a huge accuracy threat. In such cases, human analysts are hard pressed to distinguish the records of the person in question from their similarly named relatives. Producing a background check suddenly transforms into a task wildly susceptible to the prospect of misplaced data by human error.
An analyst working on the Thompson report would have searched his legal records in the counties he is associated with, finding a few speeding tickets and nothing more. The analyst also would have discovered that Thompson’s cousin of the same name was involved in many lawsuits. Most good analysts would be able to distinguish between the two. However, with the combined powers of AI and experienced analysts on the case, a bigger story was uncovered.
Amidst the many files associated with Thompson’s similarly named relatives, a case of insider trading was discovered. The intricate structure of the algorithm had the capacity to conduct an exhaustive search on all of Thompson’s related addresses, not only his associated counties, thus singling out his connection to a certain limited liability company that was referred to in the insider trading case. Having uncovered this essential information, the automated system provided the analyst with the tools necessary to identify Thompson as a party in the lawsuit, which may have otherwise been omitted from the background check. Screening individuals for potential risk requires the utmost attention to detail, alongside the ability to review copious amounts of data, a challenge which is inevitably subject to human error, but a provision suited perfectly for the use of artificial intelligence.
Confirmation Bias: Jeopardizing Accurate ODD
In an effort to work efficiently, the human mind tells itself stories. Often, without being cognizant of it, we make assumptions and fit all corresponding events into that story. This phenomenon is commonly referred to as confirmation bias. When it comes to comprehensive background checks, confirmation bias is an unavoidable psychological anomaly, even for the most seasoned research analyst.
For instance, when conducting a background check, one of the most important elements of the process is pinpointing legal records that are relevant to the subject of the search. A research analyst will comb through large amounts of data to duly match records that correspond to the individual in question. Intermittently, analysts come across lawsuits that list modified versions of a subject’s name as a party. Through additional identifying factors, an analyst can determine if the modified variations of the individual’s name is, in fact, the same individual they’re researching. If all signs point to it being the same person, analysts should re-run the modified name through all of the corresponding databases. Here’s where confirmation bias comes into play. Without being cognizant of it, analysts may often reject that this modified name is in fact an alias for the fund or portfolio manager they’re researching. In place, they will choose to stick with the information that confirms the story they have already created for the subject in question.
For example, an analyst conducting a background check on a person guilty of a few traffic infractions will naturally regard that individual as a typical, law-abiding citizen. If the analyst discovers a securities fraud charge against someone with the subject’s alias, i.e. a nickname or initials, they often conduct a brief and superficial analysis, eager to dismiss the case as irrelevant. As a result, identifying factors may be overlooked, and a criminal case falsely omitted from the background check. Confirmation bias clouds the analyst’s judgment, making it difficult to deviate from their initial impression of the subject and dig deeper to properly produce an accurate and reliable report.
This vulnerability highlights the fundamental difference between human scrutiny and automated analysis. Artificial intelligence approaches segments of data with an objective view, thus eliminating the risk of misplaced information by virtue of confirmation bias. Systems that leverage AI will dig deeper to ensure that the subject is connected to the case, due to the fact that they simply are not programmed to make these natural connections and psychological shortcuts.
Not only does AI assist in subduing psychological interferences, it also overcomes physical restraints. Due to the circumstances of human limitations, companies struggle with maximizing the work that can be accomplished within a certain time frame. AI, on the other hand, automates arduous tasks and repetitive duties, thus ensuring time is used more productively. It’s no wonder why the results from a study conducted by PwC showed that 54% of executives reported a drastic increase in productivity after implementing technologies that leverage artificial intelligence. In fact, an analysis by Deloitte suggests that an analyst working in conjunction with an AI system could save as much as 364 hours a year. For operational due diligence managers, time is of the essence as other teams are awaiting their input regarding investment decisions. Taking ODD managers’ workflow into consideration, some automated systems produce background checks before the pressure hits, at a rate seven times faster than industry standards.
Nonetheless, it is not enough for automation to expedite time-saving output if it isn’t supplemented with a higher degree of accuracy. Precision is a key factor in sustaining the competitive advantage of those operating with AI-powered systems. One of the many advantages of automated, artificial intelligence systems is the ability to sift through piles of data and scrupulously extract relevant information. When it comes to operational due diligence, this is extremely beneficial. Background checks require close attention to detail in order to avoid inaccuracies. AI’s sophisticated algorithms are designed to derive accurate data, far surpassing the proficiency of manual research alone, due to the natural human errors noted above.
New Look Ahead
Comprehension. Accuracy. Speed. These are crucial factors that are essential to conducting an effective background check. A dependable, factual and transparent ODD procedure is essential, especially one that can deliver results promptly. AI has been instrumental in improving accuracy.
Moving away from the advances AI is making in regard to the compilation of reports, there are significant strides related to the user experience when running and reviewing background checks thanks to this technology. When background checks, an essential part of the screening process, are an onerous procedure, it constitutes a setback for making well-informed decisions. Whether hiring a candidate, proceeding with an investment opportunity or evaluating a company’s management—accessible due diligence is a crucial requirement.
Bringing the power of the “one click revolution” that Amazon is attributed to introducing to the world, some background check platforms are reworking the flow of background checks, from submission to review, to create a fluid user experience. With the efficiency of the one click resolution, the arduous task of submission becomes as simple as filling out a name, state of residence, and age range.
Additionally, review is enhanced by leveraging technology to create a digital report that illustratively highlights red flags and includes links to all original sources into an effortless undertaking to ensure that managing due diligence practices is more accessible than ever before.
As we move further into the new year, it is necessary to consider the implications of the advanced precision that AI has to offer to the world of compliance. The significance of AI and machine learning and its practical implementation can refine the operational due diligence process to make more informed and timely decisions.
1Liu, Shanhong. “Artificial Intelligence Worldwide – Statistics & Facts.” Statista, 4 December, 2019, https://www.statista.com/topics/3104/artificial-intelligence-ai-worldwide