Although companies have spent millions of dollars on analytics software designed to sift through mountains of unstructured data for signs of theft and fraud, a recent flurry of high-profile hacks could drive those firms to double down on “counter-fraud” tools.
At least, that’s the hope of IBM and other tech vendors.
IBM has announced a “Smarter counter fraud” initiative, involving software that scans for “fraudulent indicators.” The platform can “learn” from previous attacks to predict new ones, as well as visualize fraud patterns. IBM claims the underlying analytics are more than capable of drawing the non-obvious connections so essential in sniffing out electronic theft. In keeping with Big Blue’s recent embrace of the cloud, these tools are offered as a service.
But sophisticated analytics only go so far—at a certain point, an actual human being needs to take a look at the data indicating possible fraud and make a decision about what to do. That’s where things get problematic: according to an extensive report by Bloomberg Businessweek, for example, company officials at Target declined to respond to data suggesting a massive breach was in progress, which eventually led to the theft of millions of customers’ personal information.
Target security personnel and law enforcement officials all confirmed to reporters working on that story that the automated fraud-detection system, known as FireEye, worked perfectly; none seemed too clear on why warnings were ignored by the company’s executives. The company later admitted to disregarding the information in a widely circulated statement to the press: “With the benefit of hindsight, we are investigating whether, if different judgments had been made, the outcome may have been different.”
Given the need to keep personal data safe in a risky environment—to retain customers, if nothing else—companies will spend more on analytics tools helpfully promoted by tech vendors who see a sizable buck to be made in fraud protection. No matter how advanced the software, however, poor human decision-making is a problem that so-called “Big Data” can’t solve.