Another massive, expensive, and painful IT failure was reported in the IEEE Spectrum of 24-JAN-2018, this time, it was the MIDAS Unemployment system in Michigan.
Not only did this system fail to deliver, its managers refused to admit that it failed. The system was to replace the legacy software used by the State of Michigan for unemployment insurance. It uses an algorithm to detect suspected fraud and had a small glitch on roll-out, it basically suspected the majority of unemployment claims a fraudulent. Oooops.
A thorough review found that from October 2013 to September 2015, MiDAS adjudicated—by algorithm alone—40,195 cases of fraud, with 85 percent of those resulting in incorrect fraud determinations. Another 22,589 cases that had some level of human interaction involved in a fraud determination found a 44 percent false fraud claim rate, which was an “improvement,” but still an incredibly poor result. Interestingly, but not surprisingly, the UIA has stubbornly refused to explain exactly why MiDAS failed so spectacularly, or why it ignored all the early warning signs that something was radically amiss.
The article also contains links to the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems which is heartening, at least someone is working on this and at best, it is the IEEE.