That one time AI killed me

By Ben Olson
Reader Staff

“O brave new world, that has such people in it!”
— William Shakespeare (or maybe it was a bot that said this. Who knows anymore?)

Journalists are used to covering demise. Whether it’s the dissolution of a beloved nonprofit organization, the passing of a notable city resident, a favorite watering hole closing shop or the ever-changing nature of a small town in the West, we’re fairly steeped in death. However, it was with some surprise that I was informed last week that, according to an AI-generated article, I died two years ago.

The only consolation came from the fact that I was informed of my own death by my best friend Zach Hagadone, who also serves as editor-in-chief for the Sandpoint Reader. When something enters Zach’s mind, he burrows deep into the topic. It doesn’t matter if he’s writing an article or just increasing his understanding for yucks — he attacks it with gusto.

In this case, Zach was exploring the limitations of the ChatGPT bot, which has made headlines the past few weeks for its ability to spit out text in seconds that it takes us dumb humans hours, even days, to produce.

Zach’s experiment was to prompt the bot: “What is the biography of Sandpoint Reader Publisher Ben Olson?” 

[Editor’s note: The bot didn’t know “specifically” who Zach Hagadone was; however, it conceded, “an editor named Zach Hagadone may work for this publication.” It therefore could not conjure a biography, which kind of bummed him out, to be honest.] 

The response it spat back was at first laughable, then a bit eerie.

“Ben Olson was the publisher of the Sandpoint Reader, a weekly newspaper serving the town of Sandpoint, in the Pacific Northwest,” the bot began. “He was born and raised in Sandpoint and had a deep love for his community and its rich history.”

I had a smile on my face while reading up until the end, when ChatGPT decided my time on this earth had come to a close.

“Olson passed away in 2021 at the age of 55,” I read, spitting out the water I was drinking. “He left behind a legacy of excellence in journalism and a deep connection to the town of Sandpoint. He will always be remembered as a dedicated publisher, journalist and friend to his community.”

Needless to say, I’m not 55 years old, nor am I dead. While I appreciate the empty praises this bot heaped on me, it’s yet another example of why we’re not quite doomed when it comes to chatbots creating journalism — they’re completely unable to access any facts or information that haven’t already been reported, written down and shared by a real live journalist with a pulse.

Regardless, my premature virtual demise got me thinking some dire thoughts.

An image generated by AI when entering the prompt, “Van Gogh sky with pigs.” Image courtesy Night Cafe.

If you don’t know much about the ChatGPT bot or artificial intelligence in general, it’s rapidly becoming the largest existential threat to artists and creators we’ve ever witnessed. Begun as a curiosity, AI has now been lauded as a breakthrough that will save time and money. In reality, AI is yet another way we are destroying humanity, one “innovation” at a time.

Visual artists are already reeling about the AI takeover of their discipline. Websites like Midjourney and Night Cafe make the arduous process of creation as simple as a few keystrokes. What hitherto took hours, days or weeks can now be accomplished in a moment after typing a few prompts into a software program — no talent necessary.

Type “Van Gogh sky with pigs” into the prompt and, three seconds later, the website kicks back an original piece that not only looks stunning, but can be tweaked to satisfy any number of parameters.

It’s so good, in fact, that AI-generated art is capable of winning awards over real artists.

When Jason Allen entered the Colorado State Art Fair’s competition in the category of “emerging digital artists” in late 2022, his entry was made using Midjourney. And it won, much to the chagrin of his competitors.

“It’s over,” Allen told The New York Times. “AI won. Humans lost.”

The award earned Allen just $300, but the blowback he spurred by beating real artists with a software prompt was enough to rattle the cages of the art world.

“This thing wants our jobs,” tweeted RJ Palmer, a movie and game-concept artist. “It’s actively anti-artist.”

Visual artists like Kelly McKernan are beginning to take matters into their own hands, filing a lawsuit against the website DeviantArt, to which she had posted her original paintings for 20 years. McKernan found that when AI-generators went online in 2021, many people were typing “in the style of Kelly McKernan” when prompting AI to create a new piece. 

“There’s more and more images with my name attached to it that I can see my hand in, but it’s not my work,” McKernan told NPR. “I’m kind of feeling violated here.”

DeviantArt allowed its users to pay a monthly subscription fee to get access to the AI art generator, trained on images from artists like McKernan. It didn’t pay the real artists a single cent, though, bringing on a lawsuit filed by McKernan alleging that DeviantArt and AI companies violated copyright law by compressing billions of images and storing the information, which it then used to make new works.

That’s the main downfall of AI-generated art (at least for now): When AI creates art, it is trained to steal the work of current working artists. 

It’s true that Pablo Picasso once said that while, “good artists borrow, great artists steal,” the emergence of AI-generated art and writing might trigger a redefinition of plagiarism in order to make it easier to create content with AI instead of humans. Why bother? Because money.

While sites like Midjourney and others are unsettling artists around the world, ChatGPT has caused the same upheaval in the world of journalism and creative writing.

Launched in November 2022, ChatGPT also started as a novelty, but has quickly grown into a threat to creative writers everywhere.

Utilizing “large language model tools,” ChatGPT can indeed produce articles and original pieces of writing — including my own obituary — but the technology isn’t exactly there yet.

There’s a saying that if you put enough monkeys and typewriters in the same room, one will eventually recreate Shakespeare. The difference between infinite monkeys and ChatGPT, however, is in the way the latter creates language.

Chatbots like GPT are powered by large amounts of data in order to string words together in a meaningful way. They not only tap into vocabulary, but also understand words in context, helping mimic speech patterns.

There are companies attempting to use AI for good, like a bot called Rentervention, aimed at supporting tenants while dealing with landlords. There’s also the DoNotPay service, a “robot lawyer” that helps resolve customer service disputes and even offers representation in courtroom traffic cases. Yet, the vast majority of AI for writing purposes is seemingly engineered to replace writers.

Many media companies are already utilizing AI for content generation. Buzzfeed recently caused a stir when it announced ChatGPT creator OpenAI would help produce quizzes and other forms of content. Instead of doom and gloom, the news actually drove a brief 200% spike to Buzzfeed’s stock price.

How does this all relate to AI killing me? Reading my own death notice, as rendered by ChatGPT, revealed some critical failings of the technology.

“Olson was a journalist by trade and had always been interested in the power of storytelling,” it “wrote.” “He started his career as a reporter for the Sandpoint Reader, quickly rising through the ranks to become an editor and eventually the paper’s publisher.

“Under Olson’s leadership, the Sandpoint Reader continued to thrive, despite the challenges posed by the digital age,” the bot continued. “He was a strong advocate for quality journalism and believed the role of the press was to hold those in power accountable and serve as a voice for the community.”

It was about this point when I realized the truth about ChatGPT. Just like those useless Buzzfeed articles, written by either a human-like bot or a bot-like human, there was a tiny bit of substance surrounded by a heap of nonsense. It’s almost like when you hear someone talking in a dream. It’s definitely language, and some words are recognizable, but most of the substance of the conversation is lost because nothing really makes sense. 

While some of the things the bot recited about me were accurate, the statements were shallow, held very little conviction and were extremely light on facts.

It was just truthful-sounding enough to pass cursory examination, but ultimately a work of fiction.

We have no business opening this Pandora’s box. Art, whether it’s a surreal painting, a detective novel or a newspaper article, is the very thing that defines humanity. Without the ability for humans to access the pathos of our dumb, beautiful world, we’re truly lost souls pissing into the abyss.

AI-generators are proving, once and for all, that the free market truly does not care about humans. In order to save a few bucks and see the stock jump, we are literally selling the last vestiges of the human experience to the highest bidder, all to satisfy the perception of growth and innovation. 

While AI isn’t currently passable without a human’s touch, as the years progress, we might just find a way to make our own existence unnecessary. That’s what scares me the most.