A.I. Doesn't Always Mean the End of the World
By Kirk White, Yusen Logistics
Continuing our Supply Chain Futurist series but segueing off our Blockchain exploration to visit with another component of the coming revolution: A.I. Sometimes referred to as “artificial intelligence” and sometimes called “augmented intelligence”, A.I. is most commonly referred to as…OH NO SKYNET IS SELF AWARE AND THE TERMINATORS ARE COMING AND THEY WON’T STOP UNTIL WE ARE ALL IN THE MATRIX…simmer down. It’s easy to get lost in the pure dystopian speculative fiction of it and it certainly doesn’t help that every single sci-fi book or movie that concerns A.I. (Haley Joel Osmet notwithstanding) usually ends up with some rag tag band of freedom fighters using homemade rockets to go…fight…win, and it is certainly of concern that our best Jeopardy and Chess minds have already fallen to our PC Overlords…but before we all hide in the desert, let’s remember that, presently, computers aren’t even able to recognize squiggly letters or discern which of these ten pictures contains an image of a car; so let’s not call John Connor just yet, okay?;
A.I., in its simplest form, is computer learning—An application with the ability to analyze anomalies and errors in performance and adjust for the next time around. Imagine a young boy. He is learning to speak but has trouble with a few words. He wants the tennis ball can interactive portal to the ether named after an ancient Greek library to play his favorite song, “raining tacos” (do NOT look this up…you will never get it out of your head). So he goes to the tennis ball can and says, “playsrainntacos” and the tennis ball can says “sorry I don’t understand” and he says again “playsrainntacos” and again the can says “sorry, I don’t understand” and so the boy’s father steps in and says “Play it’s raining tacos” and the can complies. “Raining Tacos” fills the air. And this cycle goes on a few times, because “raining tacos” is quite catchy. And then suddenly, the boy says “playsrainningtacos” and the can says “Playing Raining Tacos” and the airways once again are alive with shells, meat, lettuce and cheese cheese cheese. This is A.I. The program “learned” by making an error. Every time the request was misheard, and the adult voice provided the clearer correction, the application inside the tennis ball can took that information and made a new connection. “playsrainnintacos” = “play it’s raining tacos” and, just like a human brain (albeit way way way less complex) a new path was formed.
Not so much.
Adapting based on anomaly is more like it.
But then again, isn’t that what learning IS?
There are three stages of A.I. and we are in the very early chapter in the very first book. It breaks down like this:
Stage ONE: ANI—Artificial NARROW intelligence. We can train a computer to be an expert at ONE thing; better, faster, stronger at ONE thing. Chess computers. Poker computers. That random time some program wrote that Harry Potter Book. Analyzing a bunch of data and identifying patterns and adjusting their programs to accommodate without the need of external assistance. This is what’s happening in our tennis ball can example above. So yes…we have this NOW. This is the level of A.I. currently available in our world and it’s taken the entire history of computers to get to this point.
And we are working (but haven’t gotten there yet) towards:
Stage TWO: AGI—Artificial GENERAL Intelligence. This is where it gets a weeeeee scary. This is human level intelligence. And before you say, “well I’d say my phone is already smarter than Uncle Jerry”, it’s not. Because we’re not talking about math or memory skills here. We’re talking processing of data in CONTEXT. If you think a moment about the last time you interacted with a friend or coworker and just “knew” he/she was having a bad day. Or maybe the last time your spouse or significant other was mad at you; how did you know? Perhaps you picked up on a facial expression, a furrowed brow, a grimace, certain downward glance that elicited a “what’s wrong”. You analyzed and processed based on physical indicators that a computer couldn’t even begin to process. Computer would say “that’s a face”.Again, computers can’t even click a box that says “prove you’re not a robot”. But it’s coming…that is the next level of A.I. A computer with the ability to process based on random data sources and “intuit” based on non-logical indicators. And then we will have true Artificial Intelligence. You can even make a new friend…until…
Stage THREE: ASI—Artificial SUPER Intelligence. This is Skynet. This is when computers will surpass human processing and enter into the realm of…actually, that’s the scary part…no one really knows what that will be like. Or what the programs will be like.
And here’s the final boil to your noodle. Many believe that, even though it’s taken DECADES to get from the first computer programs to ANI, once AGI is achieved, ASI will happen like 20 minutes later. Because…computer stuff.
And then maybe the Matrix.
But for NOW… the world is beginning to see some serious plus-sides to ANI, especially in the world of supply chain. The immediate benefits are in the realm of analysis. The sheer speed at which data can be collected, examined and broken into patterns is astonishing. Imagine a report that can predict potential disruptions in your product flow AS you’re uploading the P.O. Oh, this shipment may be delayed because there is a 98% of a tropical storm happening the month it is supposed to ship based on 100 years of weather reports and there may be issues at the port because historically when these certain conditions happen, a month later there are labor disputes and oh, the history of every order from this particular vendor tells us that October is their highest month of employee turnover—so may we suggest changing factories or marking this as a hot shipment? And this is instant? And it’s more than likely on the horizon.
Other benefits include “teaching” the supply chain. An SOP could be uploaded into the system and used to build automatic logic-based variances to conditions (+/- 5% quantity is okay to ship as long as it’s not #BR549) and communicated throughout the order process to pre-screen exceptions.
The list is endless and many companies are already looking into this and it’s predicted that within one to two years, as many as half of all companies will be using some form of A.I. in their processes.
Next time, we’ll continue this journey into the future. Any topics you’d like to see covered, please leave in the comments.
Until then, follow the white rabbit, Neo…
Kirk White is a corporate creative and a supply chain futurist. He has worked in every division of Yusen Logistics. After a brief stint in Transportation, he transferred to Corporate, where he coordinated Yusen’s Employee Empowered Kaizen system and served as a Specialist for the Business Process Re-engineering group, after which he moved to the Warehouse division to serve as the East Coast Quality Manger before ultimately joining the International division, where he hopes to use his Quality knowledge base to prove an asset to OCM.