22 March, 2018 // Last week, the Wall Street Journal questioned the wisdom of an increasing blind faith in technology.
Its story, “Artificial Intelligence Rules More of Your Life. Who Rules AI?” (behind the WSJ paywall), was most mostly about ethics: If we let companies’ proprietary algorithms make business decisions, how can we trust that they’re acting honorably? For instance, can a self-taught, artificial intelligence program be called racist for denying home loans to a group of consumers?
As supply chain practitioners, we were more interested in the underlying problem with AI. It’s that the Big Data movement relies on logic that often isn’t publicly disclosed—or even understood. “Computer algorithms aren’t always clear on their logic,” the story reported. “What if the algorithm doesn’t know how to explain itself?”
The Journal cited the real-world example of “an algorithm that had learned—by itself—to distinguish between wolves and husky dogs.” Researchers discovered that the program “appeared to be doing so by noticing snow on the ground in the wolf pictures, not because of any insight into the animal.”
Yikes. It doesn’t take a human genius to see how that logic strategy could come back to bite you in the dog days of August. Literally.
That’s exactly the problem with supply chain technology. If your analytics program is making ops decisions because of the snow in the picture, it’s only a matter of time before you have a problem. And, there’s a lot of snow in current-generation supply chain analytics tools: Un-digitized nodes. Reported data that varies from what actually happened. Padded wait times.
Those little inaccuracies add up over the breadth of a full supply chain. Don’t get us wrong. At Morgan, we love data so much that we created a separate, sister company to build end-to-end visibility and analytics tools. Yet, data is dangerous without human wisdom and experience, which is why we let our supply chain experts run the data instead of letting it run them.
As with so many things, Steve Jobs may have said it best when he argued in a keynote that his company, Apple Inc., lived at “the intersection of technology and the liberal arts.” That worked well for a tech company trying to differentiate itself through better design and usability.
Morgan positions itself in a similar way in our work transforming inventory efficiency and in-transit operations. We deliver superior results that come from both superior information and superior insight.
In an industry where operations wolves at your door can devour millions of dollars, that’s an approach worth considering.