Session

In this sure to be calm, measured talk, I offer an argument against the use of LLMs in any language work as a high cost, low benefit “tool.” The argument is based on personal experience, a growing base of academic research, and an examination of the costs/benefits. I also try to demonstrate more clearly what LLMs actually do in terms of proposed use cases, output, and background. In short, I will argue that LLMs are engines that produce only “bullshit” (in the academic sense proposed by Harry G. Frankfurt) that put professional quality and credibility at risk.