The linguist Emily Bender has recently given another illustration of the way AI cannot grasp meaning. The goal was to illustrate what large language models, or LLMs — the technology behind chatbots like ChatGPT — can and cannot do. The setup is this: 

Say that A and B, both fluent speakers of English, are independently stranded on two uninhabited islands.They soon discover that previous visitors to these islands have left behind telegraphs and that they can communicate with each other via an underwater cable. A and B start happily typing messages to each other. 

Meanwhile, O, a deep-sea octopus who is unable to visit or observe the two islands, discovers a way to tap into the underwater cable and listen in on A and B’s conversations. O does not speak English (or any other language) but is very good at detecting statistical patterns. Over time, O learns to predict with great accuracy how B will respond to each of A’s utterances. 

One day, the octopus starts impersonating B and replying to A. This ruse works for a while, and A does not distinguish between the messages she receives from B and O, believing them all to have meaning and intent. Then, one day, A writes: “I’m being attacked by an angry bear. Help me figure out how to defend myself. I’ve got some sticks.” The octopus, impersonating B, obviously fails to give a helpful reply. How could it succeed? The octopus has no referents, no idea what bears or sticks are. No way to give relevant instructions, like to go grab some coconuts and rope and build a catapult. Result: A is in trouble. 

ChatGPT, like the octopus, lacks access to real-world, embodied referents. It lacks cognition and has no idea what the data it handles means. ChatGPT and similar AI engines handle word forms, not word meanings. Understanding meaning relies on qualitative factors that are irreducible regardless of how much data is provided.  

So, no, I wouldn’t ask ChatGPT for the Meaning of Life.