As you hear my voice, you might think that the words I speak come from me and are an expression of my perspective. But are they? What if they were written by ChatGPT? Would you be able to tell the difference?
The current problem with all written content is not that it may have been generated by a large language model. The problem is that there is no credible way to tell. And I know what you’re thinking, there are detection tools that can allegedly perceive the residue of AI generated content, like composing sentences with the form I just utilized: “The problem is not this; it is that.” What is really at issue here, however, is uncertainty — the uncertainty of not being able to tell whether something has or has not been AI generated.
So what do we do about this? Instead of desperately trying to distinguish real 100% organic, human-generated content from synthetic AI compositions, we might do better by following the advice that is contained in a question provided by Samuel Beckett in the story “Texts for Nothing”: “What does it matter who is speaking, someone said, what does it matter who is speaking?”
And there’s the rub. We believe that the source of the words — the one who is speaking — provides a guarantee of their value and validity. But that assumption is just that; it is an assumption. And the large language model doesn’t just call this into question. It challenges everything associated with this logocentric way of thinking.
I’m David Gunkel, and there’s a good chance that this is and yet might not be my perspective.
