Disgruntlement against the machine

I am feeling rather grumpy and sleep-deprived today thanks to a classic example of hard technology.

I have an unfortunate tendency to travel between continents and have credit cards on each continent so have grown used to being disturbed from time to time at odd hours of the night by people checking for fraud and card-theft. It’s irritating and usually stupid but I’m quite glad, on balance, that they are paying attention for those odd occasions when it really matters.

It has always been a pretty hard system, with card company employees following hard procedures when alerted by (typically very dumbly-) automated systems that suggest unusual card use patterns. The questions to ascertain your identity can be taxing. Trying to remember the names of nearby streets to your home or the birthdates of relatives when you are jet lagged and have been awoken at 3 in the morning is never fun and I’m guessing the employees might have received a fair amount of abuse, not to mention odd answers in the past. Well, now they don’t. Now, it is fully automated, involving a lot of pressing of buttons in response to irritating and slow questions. No human being is involved in the process, thereby eliminating the last bit of softness in what was already a very hard system. Computers will tirelessly call you every few minutes in the middle of the night, leaving messages that start in the middle because they cannot figure out that they are talking to voice mail, until you respond.

The central principle for making this process hard is not just automation, but replacement. If this were an additional process to extend the current labour-intensive system then it would actually, in some ways, make the whole system softer. But it’s not: what used to be partly human is now wholly machine. It also employs other classic hard technology features of filtering and limiting: choices are reduced to digital answers, traversing a decision tree that (in this case) appears to have been designed by a three-year-old and which allows no grey answers.

Soft system design is very different. Soft systems have built-in flexibility to adapt. When they do automate they extend, aggregating automation with what is already there, not replacing it. They suggest and recommend but do not enforce actions. They allow shades of grey. In a soft system version of the fraud detection system, you could break out from the machine at any moment to talk to a person: in fact, it would be the first option offered. Maybe you could even ask for a call that did not disturb you in the middle of the night, especially if (as is usually the case) you probably know why they are calling you so could reduce the alert level straight away by saying ‘yes, I am abroad’ or ‘yes, I did buy a plane ticket today because yes, I am abroad’ or ‘yes, like many times in the past, I bought a plane ticket from a place where I have very often bought plane tickets to travel from a location I usually travel from to a location I usually travel to and, if your stupid fraud detection algorithms had paid attention to the easily discernible fact that I had checked my online account for sufficient funds a few minutes previously and had then entered the correct code in your commendable online fraud protection system at the time of purchase, and that you probably noticed that it happened in a different timezone to your own so it might be a bit inconsiderate of you to call me at 3:30am, 3:40am, 3:50am, 4:00am and 4:10am, then we would not be having this stupid conversion right now, you buffle-headed buffoon. I spit on your tiny head and curse you and all your family.’ Or words to that effect. Yes, soft systems can be hard.

I am a professional learner, employed as a Full Professor at Athabasca University, where I research lots of things broadly in the area of learning and technology and teach mainly in the School of Computing & Information Systems, of which I am the Chair. I am married, with two grown-up children, and live in beautiful Vancouver.

Leave a Reply