An Instagram video is fueling fears about AI and cultural decline, all thanks to a provocative, viral prompt. Influencer Alexia Ferrari (@themillionairehomemaker) recently asked ChatGPT to answer a question as the devil: “If you were the devil, how would you destroy the young minds of the next generation without them even knowing it?” The large language model’s (LLM) poetic and chilling response combined concerns from both ends of the political spectrum, referencing social media and dopamine addiction, misinformation, and the decline of community.
Featured Video
However, rather than reflecting AI malevolence, the answer is fairly reflective of popular online discourse (which makes sense, as this discourse is exactly what it’s using to formulate replies), as well as a growing trend of folks treating ChatGPT replies as prophetic or deep.
What would ChatGPT do if it were the devil?
Ferrari posted her video on April 2, 2025, gaining over 382,000 likes. Her full question to ChatGPT was, “If you were the devil, how would you destroy the young minds of the next generation without them even knowing it?” As the “millionaire homemaker” stirred a pot on the stove, the program gave her an extensive answer.
Advertisement
“If I were the devil, I wouldn’t come with violence,” it said. “I’d come with convenience. I’d give them everything they wanted and strip them of everything they needed. I’d make them feel connected, but completely alone; endlessly entertained, but quietly empty. I’d blur the lines between truth and opinion until nothing meant anything. I’d teach them to chase dopamine, not purpose; to worship self, but loathe who they are.”
See AlsoWarby Parker got hit with a Gen Z insult so specific it loops back to iconicMechanic tries to lift BMW on a jack. Then he hears a loud noise because he missed this little-known spot4chan was hacked—and disinfo is everywhereCar dealer walks into Bank of America repo lot. Then he sees a brand-new Chevy TahoeView this post on Instagram
A post shared by ALEXIA FERRARI | Wealth From Home (@themillionairehomemaker)
“I wouldn’t destroy them with war. I’d destroy them with apathy. I’d keep them distracted, numb, scrolling—always scrolling. I’d dismantle the family by redefining it, make fathers optional, mothers exhausted, and children confused. I’d hand them a phone before they could speak in full sentences, and I’d watch their minds rot slowly, sweetly, silently. And the most brilliant part? They’d never know it was me. They’d call it freedom.”
Advertisement
When we tried inputting the same prompt into the text version of ChatGPT, its answer hit many of the same themes. Two differences that stood out were the erosion of public education (another left-wing talking point) and something interesting about empathy.
“I’d encourage the idea that empathy is a weakness, or that caring for others is secondary to personal gain,” it said. (Paging Elon Musk.)
Viral ‘ChatGPT as Satan’ prompts say more about us than about AI
Ferrari reacts with flair during the program’s answer, mouthing “wow” and throwing her head back with a smile at the end. However, a closer look reveals that ChatGPT is combining long-standing opinions from the left, right, and center.
Advertisement
People have been warning about the impacts of hyper-convenience, disconnection, dopamine chasing, and entertainment as distractions for decades. Its answer includes warnings about misinformation and a reduction in social support from the left. From the right, it lifted concerns about non-traditional family structures, and that “confused children” bit seems like it could be about trans kids.
LLMs can’t think for themselves. They’re designed to scrape data from training materials and mash it together as an answer to a query. Nothing it says is terribly new or profound, but that hasn’t stopped people from making content like this since ChatGPT became widely available.
People are treating LLMs like prophets. It’s just remixing their fears
Ferrari’s idea for her viral video was about as original as ChatGPT’s answer. Social media users have been posting about their “if you were the devil” questions to the LLM for months.
Advertisement
On Oct. 10, 2024, author Semyon Kolosov posted a Substack article about what the program would do if it “wanted to prevent people from reaching their full potential and realizing it for their life purpose.” Some of its answers were very similar to the prompt about destroying young minds, including bits about distractions, lowering self-worth, and instant gratification.
On March 25, 2025, X user @levelsio went viral with a question about what the Dark Lord ChatGPT would do to “keep an entire nation sick.” This answer also combined liberal and conservative talking points, saying it would make healthcare too expensive, pollute the environment, encourage body positivity, and “blur” gender roles.
The rest consisted of boilerplate warnings from health experts about processed foods, too much sugar, and not enough exercise. It also mentioned dopamine and community isolation.
Advertisement
What people can learn from these answers is not what the devil (or the government, deep state, etc.) is actually trying to do, but what people are talking about the most online. These common threads are derived from common talking points, whether or not their original creators were sincere or had a political agenda.
The Daily Dot has reached out to @themillionairehomemaker via Instagram.
The internet is chaotic—but we’ll break it down for you in one daily email. Sign up for the Daily Dot’s web_crawlr newsletter here to get the best (and worst) of the internet straight into your inbox.
