Posted on by

AI-Pilled People are Obsessed with Stealing Work and Credit for Profit

Reading Time: 8 minutes.
Those who can’t do, steal

Text written into static with a robot emoji. Text reads, "If you put 10,000 humans in a room, eventually they will write Shakespeare."Journalism is dying. We’ve choked it with poorly thought-out revenue streams that include expensive “all or nothing” subscriptions, easily blocked ads, and worse, dwindling reading skills and attention spans that deem any article above 900 words “way too long.” Demand for the written word has never been lower.

Sure, companies could stop trying to grow larger than their means. They could stop buying up every small publisher and closing it. They could fix their revenue streams to support people who just don’t have the money for the $100/month in subscriptions they’d need to read the news. And they could try a combination of formats to reach all audiences.

Instead, these uncreative trust fund babies in C-suite roles all do the same thing: layoffs. Without a workforce, who will write articles? AI, of course! But what if you don’t have any credibility anymore because the articles are junk written by AI? What if Google doesn’t boost posts without a human author? Slap someone’s name in the byline, even if they didn’t write the article, of course! Just borrow their identity and legitimacy. You could even have the AI try to write the piece in their “voice!” Or perhaps your few writers remaining could just use an AI editor, based on other famous writers, editors, and journalists, making it sound like them.

Sure, these AI-backed options are terrible, potentially open you up to lawsuits, piss off your writing staff, lead to union-backed walkouts, and don’t actually make the news better or worth paying for, but they do use a buzzword: AI. The unskilled simpletons using family money to become business leaders only know one thing: buzz.

Honestly, we could probably cause the collapse of multiple industries if we just jingled keys in front of the right nepo babies.

Two controversies rubbed salt in the wound of unemployed and struggling journalists over the past months. Grammarly made writing suggestions based on real people, claiming what they’d say about your writing, if given the chance. Alongside that, a media corporation of newspapers upset their journalists with bylines that made the use of AI by authors unclear, tagging journalists on pieces they didn’t directly write.

Journalism is dying, and people too unimaginative to make a better future are making it worse.

Grammarly Takes the Words out of Experts’ Mouths

Grammarly added “expert reviews,” seemingly the guided AI-assisted reviews of actual writers, journalists, editors, teachers, and other experts. Unfortunately, the advice appears to be generic AI-based writing advice attributed to people who did not consent to their inclusion in Grammarly’s feature.

AI is trained on stolen material, often from these experts who have contributed so much to the collective memory of humanity through the written word. No wonder companies continually cross the line. The technology itself is built from stolen labor: intellectual property used without consent and underpaid workers. How do you learn to draw a line when you’ve already crossed it?

Grammarly didn’t solely include people who could consent to AI tooling that claims to give writing advice in their voice. They included a number of people who lost their voice the way we all do, through death. This includes William Strunk Jr, who wrote Elements of Style, William Zinsser, who wrote On Writing Well, Carl Sagan, notable scientist and educator, and David Abulafia, an English historian who died this January. It seems no one was off-limits.

Among the living, one of the more ironic examples of Grammarly’s AI putting words in a person’s mouth was computer scientist Dr. Timnit Gebru. She was one of the authors on the oft-cited “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜” paper, which was one of the earlier papers discussing the dangers of AI, including bias, stolen content, racism, and environmental effects. She may have had to leave Google for it.

The generic-sounding advice is bad enough—and provided without consent—that a number of authors made their disgust known publicly. At least one has taken the next step. Journalist Julia Angwin has filed a lawsuit against Grammarly for allegedly using her voice without permission.

“I have worked for decades honing my skills as a writer and editor, and I am distressed to discover that a tech company is selling an imposter version of my hard-earned expertise.”

– Julia Angwin in a statement

Grammarly has since deactivated the feature and apologized. However, the CEO also undermined it by defending the idea as Grammarly becomes increasingly involved with generative AI.

If this feature sounds interesting to you, pick up William Strunk Jr.’s Elements of Style. It’s a short book, but an excellent writing companion. It’s certainly better than AI pretending to be a good author telling you to use more em dashes—though I do love them.

A Newspaper Allegedly Publishes AI Article Summaries with Real Author’s Names, Without Permission

A haunting robot sketch representing AI with the text "Fill me for I am empty" around it. The eyes, nose, mouth, and ears are cavernous into a foreboding metal shape

A byline is important to journalists. It tells readers who researched and wrote the article, who put in the work. It’s why I always include the sources and links to stories I use with the author’s name in every article. People work hard to put the news together, they deserve the credit. Researching, cross-checking, fact-checking, writing, sourcing attention-grabbing visuals, it’s a lot of hard work. The precision it takes to be a journalist is not something some convoluted summarizer or autocorrect could do.

This is why I consider it outright offensive that a newspaper publisher put the names of their journalists on bylines for “articles” written by AI.

Journalists Feel Betrayed by AI at McClatchy

“If they don’t have the ability in their contract to remove their byline, we’re going to use their name.”

– Reportedly said by a McClatchy company leader

McClatchy Media has well-known newspapers like the Sacramento Bee, Kansas City Star, and others. The company has been producing news for 168 years, and serves 30 U.S. markets. They feature the work of brilliant journalists, often represented by unions, and they’re upset with the publisher. If someone took an article I wrote, ran it through summarization software, and wrote in the byline, “Reporting by Danielle. Produced with AI assistance,” I’d be livid. I don’t use AI trash like summarizers, or any other automated writing tools. I’m proud that my voice is my own. Placing both in the byline would, in my opinion, make it seem as though I used AI to write an article.

“Reporting by [author redacted]. Produced with AI assistance.”

– An example byline produced by McClatchy’s AI for non-union newsrooms, linked in The Wrap’s article

“produced using AI based on original work by”

– An improved byline journalists say was used in a union-backed newsroom

It seems that, while authors backed by the power of unions can separate their own work from the AI versions, those who don’t have such protection have far worse phrasing that could make it seem like the author used AI to write the article. According to insiders, McClatchy executives prefer to have more control over how they add these bylines. For those who can remove their name from the work, McClatchy can still create the summary itself, producing slop from their hard work. Some newsrooms reportedly have bylines that don’t even reflect the use of AI.

The phrasing could make it look as though the author assembled slop using AI. However, it’s referring to an article someone generated as a summary or rewrite of an existing piece, potentially without the knowledge, permission, or help of the original author. This AI tool doesn’t only generate shorter articles, but sometimes completely different formats, such as listicles, something a journalist may want to avoid completely. The AI may also change the tone and chooses what details to omit. Those details can change a story’s meaning to a human audience. There’s always the threat of hallucinations and falsehoods in anything AI produces as well. This could devalue the factual nature of the journalists’ original works. Despite that, many uses of this AI still features journalist names in the byline. Authors are being asked to take these “partial bylines,” even for AI articles that they had nothing to do with.

“We don’t want the public to think we have anything to do with it. We think it’s a betrayal of the public’s trust, and it undermines our credibility, and also it’s frankly kind of insulting they’re asking us to be hacks.”

– Ariane Lange, Sacramento Bee investigative reporter and union vice chair, to The Wrap

My first reaction to the AI bylines is that it feels dangerous. Making it seem as though someone wrote a story who did not write the piece or oversee it and that piece contains factual errors or even plagiarism, now attributed to that besmirched writer, is potentially career-damaging. We have laws to protect someone’s reputation against false statements knowingly said or written. While the verbiage leaves the interpretation up to the reader, it doesn’t seem to prioritize the fact that the original author often had nothing to do with the newly generated article. What could be the motivation for the company? Apparently, the use of AI without a human source in the byline could cause Google to downrank them in search engines. Humans could also place more trust in a piece that they believe was re-written by the author of the original piece with AI. This could be why companies aren’t so quick to give up practices like this. It’s all about borrowing the legitimacy of real authors for SEO-friendly AI clickbait pieces.

What is McClatchy’s “Content Scaling Agent?”

According to reports, McClatchy is using a “Content Scaling Agent,” or CSA, based on Anthropic’s Claude. Editors can use it to create summaries for stories and shortened video scripts for scrolling addiction apps like Instagram Stories, TikTok, and YouTube Shorts. It does more than that though, it can tailor those summaries based on the audience. The reports aren’t clear on exactly how that works. The tool could potentially use different language, depending on the audience, such as swapping out larger words or more complex sentences for those the audience could more easily digest given their attention span, reading comprehension, and vocabulary. However, it’s limited to stories ranging from 200 words to a maximum of 1,500 words. This article is longer than that. A good reader could make it through in perhaps five minutes, but the average is far slower. Rather than work to improve reading comprehension, we’ve instead taken the power out of the written word.

Hello, Ray Bradbury? Yes, I’m calling because that thing you wrote about is happening.

Books and magazines with a text overlay, "Just go read."

Books, magazines, webcomics, blogs, graphic novels, those cool horror story books Adam Ellis wrote (I wish I grabbed them for the photo), it doesn’t matter what you read, just go read so there’s no need for this kind of AI.

 

CSA is a Claude-based AI tool from Elvex, who reportedly provides a number of AI tools to McClatchy media. I’m sure the journalists who have trained AI without consent and are now having their names allegedly slapped on misleading articles without their consent are overjoyed with that.

AI-powered editors are allegedly filling up the main page, burying real stories on news websites infected with this kind of AI, replacing them with the AI versions. Listicles, bullet points, and other shortened articles make up the front page, the clickbait for the AI age. I wonder how search engine optimization will evolve, or if search engines will just gleam their snippets for AI results from those AI-written summaries on the front page, never looking one link deeper to find the original, human-written story. AI providing an AI-reduced story for data ingestion for other AI. The dead internet theory receives further validation.

Shockingly, AI-Pilled People Don’t Value Consent

“I’ve written about some really tough things in my career — domestic violence, sexual assault, horrible traumas. I don’t want to have to explain to a trauma victim that they can trust me with their story, but I cannot guarantee that it won’t be fed into a glorified chatbot.”

– Ariane Lange, Sacramento Bee investigative reporter and union vice chair, to The Wrap

The core philosophy of AI-pilled people has been, “Do what we want, it won’t matter anyway.” When people downloaded a few songs, they were sued for millions. But when companies gobble up the entire internet, including those songs, to process and regurgitate them, they face no consequences. If I posted an article that says it was written by someone else to add notoriety to my blog, I’d be sued. If I put items in that article that could defame someone, I could even face libel charges. But when a corporation puts someone’s name on an AI summary that reduces the verbiage, the details, and could add false information, the most they get is an angry bunch of writers and union meetings.

A lack of consent is at the heart of AI. The data, the training, the deepfakes, the pressure to use it at work, shoving models onto our computers, it all forces its way into our lives without consent. AI’s core is getting a “no” and doing something anyway. For that alone, I hope we one day win a fight against all those who stood against humanity and consent, and rid our world of this slop for good.

Sources (Written by Real People!):