ChatGPT Starts Blocking Direct Requests To Copy an Author's Style (arstechnica.com) 39
An anonymous reader quotes a report from Ars Technica: OpenAI's ChatGPT is now refusing requests to generate text that directly mimics the style of famous authors. When asked to do so, the popular LLM instead offers a response that draws on the "broad qualities" of those authors "while remaining distinct in its own voice," for example. This morning, Ars received the following response to a test prompt asking for a story introduction in the style of Stephen King: "I can definitely write with the hallmarks of atmospheric, character-driven horror and small-town dread, but I can't write in Stephen King's exact style or closely imitate his distinctive voice. Here's an original opening that captures a similar feeling while remaining its own..."
In testing, ChatGPT generated similar dodges for other authors both living (J.K. Rowling, Amy Tan) and dead (Charles Dickens, Ernest Hemingway). An analysis published by No Latency earlier this month (PDF) found the same behavior for living authors but found ChatGPT complied with style-copying requests for deceased authors. In refusing to directly copy the "exact style" of various authors, ChatGPT offered instead to capture an overall "feeling" by incorporating some of the common features found in those authors' work. That may seem like a distinction without a real difference at first glance. But the slight alteration could be legally important as OpenAI continues to fight a number of lawsuits brought by book authors alleging large-scale copyright infringement by models trained on their work. One of those suits specifically cites ChatGPT's "uncanny ability to generate text similar to that found in copyrighted textual materials," for instance. An OpenAI spokesperson did not respond to a request for comment from Ars Technica.
In the US, copyright law generally protects only a specific expression of an idea, not the more intangible style of an author. But an AI-generated stylistic imitation could become infringing if it becomes "substantially similar" to the work of the original author. "We've never had a situation in which this personal style of individual creators could be imitated as well and as inexpensively as we now have with AI," George Washington University Law School Professor Robert Brauneis told Bloomberg Law.
In testing, ChatGPT generated similar dodges for other authors both living (J.K. Rowling, Amy Tan) and dead (Charles Dickens, Ernest Hemingway). An analysis published by No Latency earlier this month (PDF) found the same behavior for living authors but found ChatGPT complied with style-copying requests for deceased authors. In refusing to directly copy the "exact style" of various authors, ChatGPT offered instead to capture an overall "feeling" by incorporating some of the common features found in those authors' work. That may seem like a distinction without a real difference at first glance. But the slight alteration could be legally important as OpenAI continues to fight a number of lawsuits brought by book authors alleging large-scale copyright infringement by models trained on their work. One of those suits specifically cites ChatGPT's "uncanny ability to generate text similar to that found in copyrighted textual materials," for instance. An OpenAI spokesperson did not respond to a request for comment from Ars Technica.
In the US, copyright law generally protects only a specific expression of an idea, not the more intangible style of an author. But an AI-generated stylistic imitation could become infringing if it becomes "substantially similar" to the work of the original author. "We've never had a situation in which this personal style of individual creators could be imitated as well and as inexpensively as we now have with AI," George Washington University Law School Professor Robert Brauneis told Bloomberg Law.
We're just getting started! (Score:1)
Is slopocalypse a word?
Re: (Score:1)
Google's AI says yes.
Re: (Score:2)
You bet. And there are other new slop words now too! For example, here is a "sloperating system":
https://github.com/Kelsidavis/... [github.com]
To save you from having to click it, this is an AI-generated, clean-room implementation of Mac System 7 OS, for x86 platforms.
AI-Nannyism (Score:4, Insightful)
This is why the future is local LLMs, not centralized, censored commercial Nanny-AIs.
I actually had Claude tell me this one time earlier in the year... right after it had just written a story in the style of the same author. So they've been trying to make it work for some time but clearly taken a while.
Re: (Score:2)
Re: (Score:2)
Apparently you haven't played with any local LLMs yet. While they are not up to the big frontier models, they are quite amazing and can do a lot of the low-quality generative stuff that most people use AI for already.
We're getting really close to having coding models that can run on local machines (by local I mean a small cluster of 128 GB NVidia dgx spark machines or AMD 395s), and just about do what Claude Code's models can do, but do it all locally.
Re: (Score:2)
Re: (Score:2)
It’s hard to imagine. Aside from being partially able to get around corporate censorship, local LLMs are incomparably shittier.
Re:AI-Nannyism (Score:4, Interesting)
You're right that you're not going to get Kimi K3 performance, requiring roughly 600GB of VRAM to run, from a small number of 24GB consumer GPUs. You're also not going to get closed-top-model performance, because you simply don't have any way to get the weights.
If we're talking about K2.7, Deepseek R1 Distilled, Qwen 2.5, Llama 3.2 though - All of which can be run locally -The question really becomes what you expect from them. They're all strong models that know more than we (individually) do about 99% of topics. A "high-school level education" is more than enough for most humans to get by in the world. Now imagine a "high-school level" understanding across every topic. It's not doing calculus, but it can tell you what patterns to use. It's not writing Shakespeare, but it can likely recite most of his works to you from memory. It's not going to come up with a clean-room reimplementation of Windows 11, but it can check your syntax and tell you the function prototypes of every publicly-available library in every programming language without breaking a sweat.
More importantly, to the GP's point - Is a model "smart" enough to understand you're trying to abuse it necessarily better than one that just carries out your orders successfully? Case in point, lately I've found Gemini's new "watcher" blocks a good 90% of my totally legit questions before the question ever makes it to the real model. Umm... No thanks? The best AI is the one that works, not the one that gives the deepest philosophical reason for saying no.
Re: (Score:2)
Claude didn't even want to discuss details about black powder production, not quite new tech, while Gemini would have allowed me to begin production immediately and spilled EVERYTHING.
So the system-prompt contains cover-your-ass text (Score:2)
It was a dark and stormy night (Score:2)
What about Edward Bulwer-Lytton? An author so inspiring they named awards after him [wikipedia.org].
Re: (Score:3)
He also inspired the name of the British beef extract Bovril [wikipedia.org]. Perhaps LLMs will get into the product-naming game (if they haven't already).
Johnston took the -vril suffix from Edward Bulwer-Lytton's then-popular novel, The Coming Race (1871), the plot of which revolves around a superior race of people, the Vril-ya, who derive their powers from an electromagnetic substance named "Vril". Therefore, Bovril indicates great strength obtained from an ox.
Okay, so it won't imitate Hemingway (Score:2)
But what if you ask it to write like Bad Hemingway [wikipedia.org]?
Re: (Score:2)
My best BH line right there.
This ruins everything (Score:5, Funny)
If I can't generate my technical documents in the style of Hunter S Thompson, AI is useless to me.
Re: (Score:3)
Re: (Score:2)
Hahaha, "The API is indescribable." (Followed by an index of 14 Markdown documents.)
Re: (Score:2)
I will not write it, Sam-I-Am,
I will not offer colored ham!
Local AI (Score:3)
The more that centralized AI turns into crippleware, it will allow local AI to flourish. Even if open models are not as good, the mere fact they won't be crippled will make them better in many use-cases.
Re: (Score:3)
There are multiple good reasons to run local AI. Guard rails may be the most obvious, but perhaps not the most important. Other reasons include not putting your prompts in the cloud for glowies to collect, not having to deal with a model rugpull, never hitting usage limits (although they're replaced by the speed limit of your hardware), and being able to fine tune a model to not even have the guard rails in the first place. If you find something you love as much as some people loved ChatGPT 4o, you can run
The problem will be cost (Score:1)
I think we are done with seeing the cost of hardware go down over time. There are too few companies making hardware and nobody can compete with them because if you try they will just lower their prices briefly until you're out of business. You won't even be
Disclaimers don't mean anything (Score:2)
Felony, or Flattery? (Score:2)
If I ask AI to write in the style of AI (cough, Madagascar, cough), is it considered a felony or form of flattery?
Asking for a professor stranded on an intellectual island where 32 out of 35 hairless apes will..
Using AI for creative writing... (Score:2, Troll)
...is a misuse of the tech
We already have authors who do it perfectly and writing is fun.
The only reason that people want AI writing is to defraud publishing platforms.
If you use AI to write for fun, fine, but it's personal fun, like playing a video game.
Nobody wants to read or pay for your slop.
I tried ... (Score:5, Funny)
Re: (Score:2)
Did you ask it to integrate new findings?
LLM should say why (Score:2)
King Style (Score:2)
Damn, I wanted to create the next (Score:1)
epic poem in the exact style of Homer (not Simpson).
Workaround? (Score:1)
Only "famous" authors? (Score:2)
So, all the rest of us, including "up-and-coming" writers, can be copied?