An AI-generated influencer entered Alabama ‘RushTok’ — and some fans didn’t care that she was fake. The experiment, created by Olivia Moore, an investment partner at Andreessen Horowitz, ran for a week and showed how quickly viewers can spot AI content — and how many choose to ignore it anyway.
The video opens with a curly-haired, bespectacled redhead named Janie B. introducing herself from a dorm room. She describes her outfit: a pleated skirt from Abercrombie, a knit top from J. Crew, socks from Target and sneakers from Nike. It’s her first video on her TikTok account, @jani.b086, posted on Aug. 8, and it has since drawn over 120,000 views.
Janie B. Is Not Real
Janie isn’t like the other RushTok girls documenting their Bama Rush experience this year. She doesn’t attend the University of Alabama, didn’t rush a sorority and didn’t buy any of those clothes. That’s because Janie isn’t real.
Moore created Janie as an experiment to understand how consumers react to AI-generated content. She spent $100 in credits producing a week of videos of her fake influencer participating in Bama Rush, then watched as her audience argued whether Janie was real or not.
How Moore Built Janie
Moore used ChatGPT and MiniMax H3 to build the character. She found a real video of a University of Alabama dorm room, uploaded it to ChatGPT, and prompted Janie to have frizzy red hair, glasses and a 1960s style. ChatGPT produced a still image of Janie in the dorm room, and the “RushTok” influencer was born.
For each TikTok, Moore fed the still dorm room image into MiniMax H3 to animate Janie. She used ChatGPT to shape how Janie moved, sounded and spoke, producing prompts that drove the videos.
The Reaction Split
Reactions to Janie were split from the start. Moore said anyone under 25 immediately spotted her as AI. But older viewers, which make up much of the Bama Rush audience, were fooled.
“I would say anyone under maybe 25 immediately clocked her as AI,” Moore said. “But anyone in the older demographic, which is much of the Bama Rush audience, was, I would say, fooled by her and just wishing her luck and sending her good thoughts in the process.”
Fans Who Didn’t Care
The audience caught on by day two. People took to Reddit to discuss Janie, made commentary videos and corrected anyone who thought she was real in her comment section. But Moore found that even after the reveal, some viewers kept watching.
“By day two, day three, the comments were filled with still some people who were saying really nice things. And people would comment and tell them, ‘Hey, this is AI.’ And they’d be like, ‘I don’t care. She seems so sweet. Like I want her to succeed.'”
Dropping Clues
As the audience caught on, Moore changed her strategy. She started dropping hints in the videos and watched as viewers pointed out clues in the comment section. She said the experiment showed how willing some viewers were to keep watching even after they knew Janie wasn’t real.
Moore’s Takeaway
Moore said humans are getting better at detecting AI quicker than AI generation is improving. She described two competing lines: how fast AI approaches true human-like quality, and how fast humans get better at spotting it.
“We’re kind of in a graph with two competing lines of different slopes,” Moore said. “So one of them is like, how fast is AI approaching true human-like quality? And that has gotten a lot better over the past year. The other line is how fast are humans getting better at detecting AI?”
The Experiment’s Point
Moore created Janie to answer philosophical questions. What do people feel about AI content? What do they feel when they don’t know it’s AI? What do they feel when something they liked turned out to be AI?
The answers came fast. The younger viewers saw through Janie instantly. The older viewers didn’t. And a good number of them didn’t care once told.
That’s the strange part of the story. Janie was a fraud from the start, and the people who believed in her were lied to. Yet they kept watching, kept cheering, kept wishing her luck — and said they didn’t care that she was fake.
It’s a small experiment, but it points at something larger about how people consume media today. The audience is not passive. They see the seams and they ignore them anyway.
The experiment also shows how quickly detection improves. Moore said viewers under 25 spotted Janie as AI immediately. Older viewers were fooled, but the pattern held: once the truth came out, the illusion collapsed.
The broader lesson is that AI detection is a two-sided race. As generation gets better, so does detection. But the gap between the two matters, and it matters now.
Moore’s experiment was a small test with a large audience. It showed that audiences can spot AI quickly, but that spotting it doesn’t stop them from enjoying it. That’s a strange dynamic, and it’s one that brands and creators will need to pay attention to as AI tools improve.
The experiment ended with Janie exposed, but the pattern held. Audiences caught on, but a good number of them kept watching anyway. The comment section proved that people will engage with AI content even when they know it’s not real, as long as it entertains them.
| Day | Event |
|---|---|
| Day 1 | Janie’s OOTD video debuts |
| Day 2 | Viewers start calling out AI |
| Day 3 | Comments filled with fans ignoring the reveal |
See the video the story is built around at Fox News.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.