The article critiques the concept of AI-generated text and the effectiveness of watermarking for proving AI involvement, arguing that it may not capture human contribution and highlights ambiguity in defining AI slop.
Before you answer, what does all written by AI even mean? Did I give Claude a six-word prompt and paste the result? Did I spend an hour arguing the point and ask it to tidy the punctuation? Did it rewrite one awkward sentence? Maybe it changed one comma. How much AI is enough before the whole thing becomes AI slop? And who decides the value by counting words? I could supply one insight and let a model turn it into 2,000 words. That single insight might be the only bit worth anything. It might make millions. It might save the planet, while the other 1,999 words are just very competent wallpaper. A detector can score the token patterns. I don’t see how it finds the idea. The reverse annoys me just as much. I could take a fully generated essay, change a comma and call it human. The comma might be meaningless. Or it might reverse the meaning of the sentence. I think all the mystical chat about AI provenance is asking the watermark to do something it can’t. It might show that a particular run of words passed through a particular model. Fine. It gives me a receipt from the last machine in the chain. One common version of the machinery can be simplified to: r_t = H(secret key, recent tokens) Where’s my idea in that equation? If we’re encoding Claude provenance into text, when do we start encoding Bob Marley provenance? If I write the argument, feed it into a model and ask for cleaner punctuation, why does the machine get a hidden signature while my contribution becomes invisible? Take Chaucer. Feed one of his poems into an AI, modernise some vocabulary and emit a watermarked version. Congratulations: the token pattern now says AI. Apparently Chaucer was merely assisting. I’d prefer hidden characters. Then I’d know I was looking at a removable label attached to one copy, and one that’s easy enough to strip. So, back to the title. This text was all written by AI. Can you prove it? Can you prove what all means? \Some of the words have been swapped to avoid Anthropic’s entirely hypothetical watermarking technology.*
An opinion piece discusses how labeling real art as AI-generated leads to false critiques, and argues that 'slop' is defined by lack of human story, not AI involvement.
The author describes using AI to generate questions and a deliberately poor first draft as a technique to overcome writer's block and memory issues, resulting in faster writing without any AI-written text in the final product.
The article argues that the 'AI slop' debate conflates authorship, productivity, and engineering quality, and proposes treating generative coding systems as high-throughput, error-prone producers within an engineering control loop, shifting scarce skills toward specification, verification, and accountability.
A philosophical critique of AI-generated content ('slop') in aesthetics, arguing that quality requires coherent intentionality and style beyond mere automatic generation.