This predates the ai bubble. There used to be a really common “plagiarism detector” (something like CheckMeIn?] that would generate a “similarity score” with a database of literature. Institutions were welcome to set their own thresholds of what they considered too similar. I hit the threshold multiple times in completely original works by using language that was simply too literary or formal in nature.

This is because all art are forms of remixing, whether it’s intentional or not. We’re teaching the wrong lessons here.

For many many centuries, art and artists, whether it’s musicians, artists, actors, writers, essayists, whoever, they have been facing an uphill battle of oversaturation in each creative industry. It’s only gotten worse in the past 50-75 years, and we’re more exposed to the sheer numbers now. We are throwing a drop of water into an ocean and hoping people will notice.

Trying to use “plagiarism detectors” against databases of millions or billions of pages is about as pointless as accusing songwriters of plagiarizing songs based on four notes. There are only so many musically-useful combinations of four notes, and they have all been used. Adam Neely has been reporting on this garbage for years.

LLMs are just making the problem even more obvious: creativity is not unique, it is not unique to people, and people have been mentally trained to expect uniqueness so much that we purposely ignore 99.999% of the material that is offered to us. As such, only 0.0001% of the ones who create earn any sort of popularity, and the rest starve to death. We ourselves are starved for content, as we consume anything that fits our extremely narrow definition of creativity like the voracious vampires we are.