DIGGING DEEPER
Will machines ever live? Instinctually, without analysis, my brain says “of course,” but that answer is accompanied by a deep uneasiness. I clearly haven’t convinced even myself! How can one dig deeper to mine nuggets of justification for such a bold claim? One method, divide and conquer, divides a primary question into a sequence of smaller, more manageable ones and then consolidates the results. As the preceding article implies, establishing a compelling definition for “life” is the natural first step in the process. For brevity, we will bypass the second step and simply accept the common definition of machine: “an apparatus consisting of interrelated parts with separate functions, used in the performance of some kind of work.” The final step, simply stated, is predicting the future of machines. A daunting task given their complexity and rapid advancement. My personal journey digging deeper into “will machines ever live?” is summarized below; let it guide others in navigating the process.
What is meant by a “life?” When exploring definitions of “life,” one encounters a broad range of candidates; individually, none of them are satisfactory for this discussion. Familiar definitions of “life” combine considerations from four basic categories: what life is made from, how life functions, genealogy, and/or special (perhaps spiritual) aspects. Traditionally, life is described as being made from (1) an organization of different parts that is (2) composed of cells which are (3) themselves built from certain biological molecules. By definition, machines “consist of interrelated parts,” but they rarely involve cell-like structures or biological molecules. They could, but why would they? Machines outperform humans at some tasks precisely because they aren’t composed of cells and biological materials! By traditional definitions, it is unlikely that machines will ever “live.” Perhaps the traditional definition should be broadened.
How does life function? As the preceding article describes, life is commonly defined as an entity that sustains itself in diverse environments (metabolism, homeostasis, respond) and may produce future generations (growth, reproduction, evolution). Luckily, more challenging questions such as the meaning of life (ultimate function) aren’t included! By comparison, purpose is fundamental in the definition of machines. The usual function list treats life as simply one incredibly complex molecule (or group of molecules) that persists and make copies of itself. Future machines will likely meet (and exceed) that very low bar (they even have a purpose)! The preceding article mentions two additional considerations (forming relationships and producing lasting effects). Both are fundamental aspects of human/advanced life, but are not present in simple life. Perhaps machines will end up somewhere in between?
Is that all life is? If machines could painstakingly print out humans, molecule by molecule, would “natural” humans consider the “printouts” human or alive? Such questions encourage our inclusion of lineage in the definition. Similarly, if machines are “brought to life” in a factory, they haven’t undergone traditional experiences of biological life. Are they living? A challenging question. The laziest answer, chosen here, is to apply “life” in the most general sense; allowing machines to “live,” but also defining another subcategory of life, nonbiological. Others may include additional requirements. Two additional considerations were included in the preceding article; irreplicable and irreplaceable. Indeed, no two humans will ever be exactly alike, but no two machines will either. However, differences in humans are (often) accepted while any differences in machines are unintended. Similarly, irreplaceability requires a subtle understanding. No biological life lasts forever. We will be physically replaced. But our physical lineage and impacts endure; they can’t be replaced. Will machines achieve that?
Working definition of life. A self-sustaining, physical entity (composition, experiences, and relations) that typically endures/replicates. That’s my personal working definition. Along with a subdivision for nonbiological life.
What is the future for machines? The near-term future can be predicted given the current status and trends. Predicting the ultimate future of machines is well beyond this discussion. Generally, machines are not made from cell-like elements or biological materials. They are not self-sufficient and can’t replicate. They don’t currently meet my threshold for being “alive.” Will that endure? A relevant trend for machines is the development of artificial general intelligence (AGI). Predicted to occur within several decades, AGI will allow machines to solve general problems, like humans. With that capability, it is reasonable to imagine robots that can take care of themselves and make more robots, self-sufficiency and replication. I predict, upon the development of AGI, that machines will meet my definition for nonbiological life.
So, maybe I cheated by redefining “life” for an answer that makes me slightly less uneasy. Others may go through the same process for themselves and come up with different answers. That is the point. Humanity is at its best when we combine all of our own answers. Let me know what you come up with. As for me, I need to continue digging deeper.
Written by Marcus Young.
