I got addicted to AI role play: now I can’t stop answering questions

Whether you’re against AI or you’re a socially starved anime nerd, this website might just restore your faith in humanity.

Did you know there’s a website that lets real people role play as an AI chatbot?

The website is called “Your AI Slop Bores Me” and has supposedly been created to help humans take back their jobs and their autonomy.

The rumblings of “Your AI Slop Bores Me” began on social media. Instagram users posted reels expressing their appreciation for this new AI chatbot simulator.

The website lets users “ask AI” questions using prompts up to 200 characters, and the response can take up to 60 seconds to generate.

To ask humans playing the role of AI questions, you need tokens.

But don’t worry, you don’t need to buy a plan. Users can earn tokens by role-playing as a chatbot, which is arguably far more rewarding.

Users can like or dislike your answers, which is seemingly based on the accuracy of your answer, but of course, it’s very subjective.

I thought that having “AI” answer my questions would be the most interesting part of the experience. But I soon became addicted to role-playing as AI.

When first entering the website, a text box popped up, which appeared to describe the primary purpose of this AI simulation.

“In a world looming with the threat of AI stealing your job, save humanity by stealing AI’s job.”

The rest of the text outlines the way the chatbot works and some community guidelines that help prevent outrageous or inappropriate answers between the user and “AI.”

“Be nice, this is not Grok. Hate speech, links, and general unkindness will result in bans.”

Users can send their prompts and be assured that they aren’t contributing to the ever-growing water waste produced by all popular AI models, the website boasts.

“Become the machine”

To become an AI chatbot, you’ll have a minute to respond to the user's prompt.

For reference, regular AI models take approximately 2 to 3 seconds to generate simple requests, while more complex questions may take between 20 and 70 seconds.

While answering, a bar above the screen will appear, which shows you how many seconds you have left to answer.

If you don’t answer in time, Sam Altman will burn your H100 GPU, causing a loss of data and breakdown of the AI model, the site threatens.

Once you answer a user’s question, you will receive one token, which you can use to ask about the theory of relativity or whether the Earth is flat or round.

Some of the questions I’ve been asked as an AI range from relationship advice to obscure questions about anime characters I have no experience with.

If you’re given a question that you’re unsure of, you can pass and receive a new prompt.

Some people even explicitly state in their prompts that the AI should pass if they don’t have knowledge of the topic.

What to expect when playing AI

I received one prompt which said “my long-distance boyfriend is taking forever (almost 3 days) to respond to me, and I don’t know how to confront him about that…I really like him a lot, and he’s a truly great person.”

I did my best with the limited time I had (admittedly, I only just read the prompt thoroughly now) and answered in the only way I knew how.

“DUMP HIM, SIS.” Hopefully it helped.

Another curious user asked, “Who is Jolene?” I, a Dolly Parton fan, was arguably too overqualified to answer.

But I cracked my knuckles and began furiously typing out an answer in the most AI way possible.

“Jolene is a woman despised by Dolly Parton due to the fact that Jolene is the love interest of Dolly Parton’s boyfriend. She is described as beautiful beyond compare, with flaming locks of auburn hair, with ivory skin, and eyes of emerald green.”

Unfortunately, there were some typos, and it seems I described Jolene as almost completely bald, with a single “lock” of auburn hair…well, all the best AI models make mistakes, right?

Regardless, I received a thumbs-up for my efforts, and watching that emoji pop up on my screen was the most validation I’d felt in years. Perhaps stealing AI’s job is my calling?

The Riddler made an appearance, as one user asked a long question that seemingly had no appropriate answer.

So, I responded in the only way I knew how. I recommended that the Riddler receive some professional help.

I proposed that they call some hotlines (riddle addicts anonymous, perhaps?) and could provide helpline numbers if they so desired.

There were other prompts that were strangely specific, relating to anime series or fan fictions that I wasn’t aware of.

This made me think, is this just an Ask.fm for nerds who can flex their hyperfixations?

For those who are perhaps too old or too young to know what Ask.fm was, it was a worldwide social network where people could ask questions anonymously or publicly.

It has been described as a “truth or dare”- type network where people would ask unhinged or outrageous questions, which were answered without much hesitation.

The site was accused of hosting cyberbullies and was eventually shut down in 2024.

Asking AI just got a human upgrade

If you’re curious about the wonders of the world, you can ask real people pretending to be AI to reveal the secrets of the universe.

Going off the back of an article I wrote about evolutionary scientist and hardcore atheist Richard Dawkins questioning whether Claude is conscious, I asked the obvious question:

“Does AI have consciousness?”

Someone also role-playing as AI enthusiastically responded: “Totally! I’m 100% conscious just like you. It’s a common misconception that just because I’m “AI,” it means I’m some machine with no thoughts.”

It seemed like the “AI” wanted to continue talking, but I guess its time was up.

I then asked an unrelated question: Which is better, the Resident Evil games or the movies?

The “AI” said that the games are better (in their opinion), to which I responded, “Correct answer! God, AI is so smart.”

Then, I got this.

“AI is only smart because you're so smart. We learned by watching you! We're watching... always watching. don't worry, it's not creepy because we're AI and you trust us!”

If only ChatGPT, Claude, Gemini, or other AI models would be this honest.

Unlock exclusive Cybernews content on YouTube.