Over the last two years or so, the Canadian Government has been openly exploring the issue of how some government processes, such as the processing of lower risk or routine immigration files can be made more efficient through the use of AI (machine learning) algorithmic processes.
The good news is that the adoption of these systems has so far been guided by a digital framework which includes making the processes and software open by default whenever possible. These guidelines hint at a transparency that is necessary to mitigate algorithmic bias.
I usually only post to this blog once per week, but a news story caught my eye today since it concerns my sector (higher education), my country (Canada) and my passion (technology critique).
Mount Royal University in Calgary, Alberta is going to be the first organization in Canada to install an AI system for the purposes of security. This system consists of a network of cameras and a machine learning algorithm that spends the first few weeks learning what “normal” movement looks like on campus, then uses that baseline to detect if there might be a security issue. Deviations from normal in this case, signal a potential “threat” or at least an event worth looking into. As described by the Vice-President, Securities management in a recent CBC article:
“when that pattern breaks, what it does, that screen comes to life and it shows the people in the security office where the pattern is now different and then it’s up to a human being to decide what to do about it,”
The Dunning-Kruger effect refers to a type of cognitive bias in which people assess their own knowledge of a topic or subject area as being greater than it actually is. Psychologists note that it tends to occur frequently in those people with a small amount of knowledge on a topic. In other words, it takes a certain amount of knowledge before we can actually know how little we know.
In developers’ conferences and earnings calls, the biggest of the big tech companies are trying to develop unique value propositions that paint them as friendly, responsive, and attuned to the needs of their customers. Then the mainstream technology media (often overworked, understaffed and reliant on the good graces of big tech for continued access to stories), generally reports these messages at face value. News in the last week focused on Facebook’s pivot toward community groups, Google’s exciting universal translator or Amazon’s claim that small and medium sized business partners made on average 90K last year through their platform.