Singularity: Shape-Shifting Self-Aware Supply Chains?
Recent AI advances indicate that self-aware, or cognitive supply chains, may be closer than we think.
Defining the terms in that headline requires a detour to an imaginary future:
Singularity—The idea that there will be a point certain when Artificial Intelligence (AI) abruptly triggers "runaway technological change" without any human input.
Shape-shifting—A being able to transform its form, shape, and especially abilities, to perform tasks outside of its capabilities or to face a new challenge.
Self-aware—When a machine begins to think for itself without programming, cognition. In sci-fi terms, it "comes alive."
Recent AI advances indicate that self-aware, or cognitive supply chains, may be closer than we think. Managing those supply chains will require not only a good grasp of best practices, but also these skills:
Persuasiveness. Great supply chain managers motivate the team to change, stay agile, perform to customer expectations, nurture vendors, and schmooze customers. Could a shape-shifting self-aware AI supply chain do that? Maybe. An AI system with a voice recently came close to beating a world-class debater in a competition. The machine listened, gathered facts, made a case, and met the human's points with strong rebuttals. The machine lost the debate, for now.
Improv. SC managers need to improvise quickly every day to deal with innumerable variables. Here's a new development. An AI-driven computer won at chess without pre-programmed gambits from past masters. AI had only the rules of the game and it improvised strategy and moves on the fly.
Even more complex than chess is the Chinese game "Go." AI player AlphaGoZero became a Go master in just three days by playing 4.9 million games against itself and learning from bad moves. "In a short time, AlphaGoZero understood all the Go knowledge that humans have accumulated over thousands of years," says Google researcher David Silver. Just tell AI the business rules, er, the rules of the game, and get out of the way.
Learn from mistakes. We call that experience and it looks like AI has it too.
Continuous education. Instead of programming in the real world, a robot dog called ANYmal quickly learns tricks by working out new actions in the virtual world. Because the virtual training regimen was sped up 1,000 times, results were achieved quickly and inexpensively and can be taught to thousands of ANYmals at one time. Just like supply chain courseware, right?
Did I say imaginary? At what point will all these pieces fit together and AI will wake up and do your job? The next 10 years? A far-off AI nightmare? No one can or will stop "progress." More likely is that as supply chains become more self-aware or cognitive, and the world truly becomes your warehouse, so too will we evolve. Convergent evolution between humans and AI is one way.