Adventures in Algorithms: My Word Against the ATM

            Once upon a time, when we regularly used CD and DVD disc drives, most of us occasionally experienced the frustration of not getting a drive to open, or close.  These devices use springs and other moving parts that sometimes do not work seamlessly.  Typically, drive failure does not result in a calamity and with some luck we can get the device to work.

            It may surprise you that the Automated Teller Machines have similar moving parts and springs, particularly where they dispense cash, accept checks and issue receipts.  I still use ATMs frequently as I like cash transactions, even though my research of credit card platforms informs me that I am subsidizing card users.  ATMs rarely fail and the most likely culprit is insufficient cash to dispense.

            Imagine my surprise when an ATM cash dispenser did not open to disperse the cash I requested: lots of beeps, followed by confiscation of my ATM card.  Lucky for me, the ATM machine failed during banker’s hours.  I made my way to a live teller, but she had little good news for me.

            In a nutshell, the ATM machine documented my transaction as completed with no problem. 

Accordingly, I have to “dispute” the debit of funds from my bank account, because the burden of proof lies with me. It’s my word versus the machine’s documentation. I will recover my $300 if and only if a manual, human accounting of transactions shows a $300 surplus.  Then and only then, will a bank manager accept my work and credit my account.  Any anomaly, like the absence of a correct surplus amount, and I am out the money.

            A non-negotiable contract governs ATM users and the transactions they generate.  I am sure that somewhere in this unread document, I have agreed that the ATM documentation trumps my word, I have no redress in court and the bank—or more accurately-- the bank’s software will establish the definitive outcome.  Put another way, I am a liar until proven innocent.

            Wish me luck.

The Internet of Infallible Algorithms

            Capital One, a major credit card issuer, offers customers access to a streamlined credit report.  Upon examining my score, I found a B rating for oldest line of credit in light of having had a card only for 12 years.  One needs 15 years for an A rating.

            Hmmmm.  My Bank of America credit card states “Cardholder since 1997,” yet repeated efforts to get that company to correct its mistake have failed.  I keep getting canned responses from—get this—a collection agency that has expended its service wingspan to providing unhelpful “answers” to bank algorithm failures.

            Bank of America has violated the Fair Credit Reporting Act by refusing to correct a credit mistake within 30 days of notification.  The algorithm lists my start date as 2003 and that becomes the truth notwithstanding what my card says, or the verifiable truth of the matter.

            So the next time a credit card company has to issue you a new card due to a security breech understand that the clock starts at zero in terms of line of credit vintage.

            In other words banks and their algorithms win every time.  They are infallible and there is no way to correct a mistake.

            This should trouble you as the future promises more algorithms and things making decisions and determining “facts.”  In my case, my credit rating takes a hit based on a clearly wrong calculation of time.

            It’s a matter of time before you find yourself immersed in a dispute that you cannot win, because the algorithms knows all and serves as judge, jury and executioner.