Emma Stamm Reported On What “Reasoning” Means In The Context of AI

For Communications of the Association for Computing Machinery, Emma Stamm wrote an article that “explores research on what’s really going on under the hood of AI models that can ‘reason’ — i.e., models that display the steps that lead to their final answer.”

Stamm told Study Hall,  “I originally pitched this as a piece critiquing the very notion of artificial reason: I wanted to focus on the possibility that there’s no true link between AI ‘chains of thought’ and the final answer, which is a hypothesis borne out by empirical evidence. My editor suggested I take a less interpretive approach. In the final version, the critical angle is there, but it’s one among multiple observations and perspectives I explore on the topic of AI reasoning. I have an academic background in this subject area, and I think this is a good introduction for non-experts that might want something slightly more technical than what you’d find in most popular tech publications.”