Credit Where Credit Is Due

One of the big fears people have with ChatGPT and its ilk is that people will take credit for writing things that were actually written by an AI. Often the talk centers around plagiarism (although I don’t think that’s the right word for it), but I’m more interested in assigning credit where it’s due in the legitimate cases of AI use.

Eric Schwitzgebel recently put out a blog post over at The Splintered Mind on this topic. Specifically, he focuses on the difference between autocomplete and ChatGPT. Schwitzgebel says that a person can legitimately use autocomplete and still take credit for their writing, while they cannot take credit for a ChatGPT output. This is because autocomplete works one word at a time and the user has the opportunity to endorse or reject each suggestion. In the ChatGPT case, a complete draft is generated, so there’s no telling if it accurately captures the user’s thoughts and intentions. That’s oversimplifying, and I recommend clicking the link above and reading the full post.

I tend towards a different approach for assigning credit for a piece of writing (or any AI generated output). There’s more to writing than simply conveying information. Writing is not just a process of moving data from one brain to another. There’s style and voice and tone to consider. I don’t think it’s enough to capture the meaning that the user intends. I think it has to be in the user’s voice for the user to get credit.

Now, I know that’s an awfully high bar to get over. It might be more accurate to say full credit instead of just credit. One of Schwitzgebel’s commentors suggested that it may be better to view the credit given for a piece of writing as a spectrum. I’m inclined to agree with that. So, in order to get full credit for a piece of writing, it has to have the author’s intended meaning, voice, tone, and style. If it has none of those things, it is not the “author’s” writing at all and no credit should be given. Everything else is an in between case. The more the author leans on the AI, the less credit is deserved.

This may seem obvious, but it’s worth stating explicitly. An author can only get full-full credit if no writing assistants are used at all. Almost no one ever gets full-full credit. Most people use dictionaries, thesauruses, spelling and grammar checkers, and the like. There’s nothing wrong with that. It just means that those tools deserve at least partial credit for the finished work.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.