“The AI Mirror: How to Reclaim Our Humanity in the Age of Machine Thinking” by Shannon Vallor will be the subject of the November First Thursday Book Discussion at noon on November 5 in the Columbia Public library.
As an award-winning professor from The University of Edinburgh with a joint appointment in philosophy and the Edinburg Futures Institute, Vallor specializes in researching the ethics of technology, particularly AI and robotics. The philosophical approach Vallor brings to “The AI Mirror” is surprisingly practical, grounded in real-life examples and outcomes.
In a relatively brief 225 pages, Vallor teases out the limits, abilities, liabilities and pitfalls of the large language models that make up the most well-known AI systems today. These models, based on statistical analysis of huge quantities of data are great at crunching through large statistical problems. Vallor explains how Google’s AI, DeepMind, has predicted protein folding structures for all known protein molecules, a computational task that had bottle-necked progress on protein research.
But the large language models lack actual experience or the ability to subscribe meaning, “Intelligence in a being that has no world to experience is like sound in a vacuum. It’s impossible because there’s no place for it to be,” Vallor writes. Which is why tasks that involve operating in the physical world — like self-driving cars, distinguishing truth from fiction, or applying morals — are still out of reach for AI.
Using tangible examples and plain language, Vallor sheds light on what these models can and can’t do, how they do it and why some tasks still elude them. At times every new page or even paragraph reveal another alarming example of AI amplifying our worst characteristics or failing to see a common sense answer with disastrous results. Vallor raises big questions. How do we proceed? What limits or expectations should we place on AI? Vallor’s analysis and reasoning gives us a solid understanding on which to project these questions and make our own meaning — something AI can’t do.
Read “The AI Mirror: How to Reclaim Our Humanity in the Age of Machine Thinking” and bring your queries, experiences and sense of meaning to our discussion.
