To get the most out of AI, we have to know what it is good at and how to use it. It’s a wonder of our modern age, and every day there’s a dazzling new accomplishment in the news. Yet its dazzle can distract, blinding us to our own strengths, leading us to surrender too quickly our own formidable powers. Knowing what AI can do masterfully and where it fails miserably is key to setting expectations and maximizing its potential.
The first is that when the BCG consultants used AI for knowledge-intensive tasks for which technology was expected to excel, labeled “inside the frontier”, it dazzled as anticipated: they got 12% more done, 25% faster, at 40% higher quality. Masterful! The second is that it performed miserably compared to humans at tasks outside the frontier, leading to a 19% performance drop, when many of the consultants blindly trusted AI answers over their own. And third, the frontier is not clearly defined. We must navigate when to use AI carefully.
AI’s strengths are savant-like, deep but not always broad. A reductive explanation of AI is that it is a pattern-finding machine. Feed it enough data and it learns patterns in that data that enable it to predict future behavior. If a forecast merely projects forward sales history, AI can predict demand even more precisely by layering additional “signals,” such as the impact an upcoming concert is likely to have on burger sales for a restaurant. AI can analyze so much data so quickly that it can recognize patterns beyond the scope of our cognitive capacity, like part combinations that are regularly late from suppliers.
That same pattern-finding machine is at play with large language models, which I call probabilistic sentence completion machines. As much as you may think it understands what you ask, the words it generates are the result of predicting the next most likely word, based on learned patterns. Generative AI responds with such finesse and aplomb that it may seem to have a personality, but underneath the hood mathematical models are generating its response.
AI is a master at patterns but doesn’t understand meaning, whereas we humans are meaning-making machines. Viktor Frankl, the famed Austrian Holocaust survivor who wrote Man’s Search for Meaning, founded a whole school of psychotherapy that posited that this drive is our central motivational force. Meaning is behind the three C’s I argue that AI lacks: context, collaboration, and conscience. AI cannot derive meaning from context, collaborate to build relationships to get things done, or make an appeal to our conscience, as Lincoln’s speech did.
AI is a transformative, disruptive force, in part because our meaning-making mentality feels uneasy with AI’s prowess. I facilitated a series of conversations with executives this past week for a forum on supply chain, and I resonated with one CPG leader, who said her aim is to simultaneously build excitement for AI but also to manage it. According to Zero100, mentions of AI on earnings calls have increased 120% over the last two years, indicating that excitement is high among boards and leaders.
A leading AI researcher and computer science professor recently lamented to me that people need to turn away from their screens and talk to each other more, because that’s where innovation happens, in conversations that spark collaboration and ideas. Google made millions from the pattern-finding of its page-rank algorithm, which returns the most relevant web pages in a search, in part based on those most viewed. But novel solutions emerge from collaboration along the paths less traveled. Generative AI is the direct result of the pioneering academic research of AI godfathers Geoff Hinton, Yoshua Bengio, and Yann LeCun, who persisted for years in their very unpopular belief that neural networks held promise. Their breakthroughs eventually proved they were right and launched an AI revolution from their very human collaboration.
Lincoln understood the need to make people powerful. On the brink of the Civil War, the nation was even more divided then than it is today, so Lincoln appointed a key rival to his cabinet, William Henry Seward, and asked for his help with the speech, which he felt was his best last-ditch effort to avoid war. Seward encouraged the president to call for unity and gave him the foundation for many of the final lines, including the angelic reference. But it was Lincoln who refined Seward’s suggestion and gifted us with that most-famous line invoking the “better angels of our nature.”
The next inaugural address by a US president will undoubtedly be written with the help of generative AI, unlike Lincoln’s famous speech. Yet mellifluous words alone are not enough. The context, collaboration and conscience behind that famous line is even more remarkable and more human. Navigating the jagged technological frontier means leveraging AI when and where we can but not surrendering the power of our humanity. Orchestrating the combination is what is needed to achieve the best harmony.
Mitchell-Guthrie has an MBA from the Kenan-Flagler Business School of the University of North Carolina at Chapel Hill, where she also received her BA in political science as a Morehead Scholar. She has been active in many roles within INFORMS (the Institute for Operations Research and Management Sciences), including serving as the chair and vice chair of the Analytics Certification Board and secretary of the Analytics Society.
The post Balancing Technology and Humanity in the Age of AI appeared first on Logistics Viewpoints.