Viral Reddit Whistleblower Post on Food Delivery App Exposed as AI Hoax
A recent Reddit post claiming to reveal internal fraud at a major food delivery app has been confirmed as entirely fabricated, highlighting the growing challenge of distinguishing truth from AI-generated content online.
Alleged Whistleblower Sparks Outrage
The post, authored by a user claiming to be a whistleblower, alleged that the company was exploiting both drivers and customers. “You guys always suspect the algorithms are rigged against you, but the reality is actually so much more depressing than the conspiracy theories,” the user wrote.
According to the supposed insider, they had been drunk and using public Wi-Fi at a library while typing a detailed exposé about how the company allegedly exploited legal loopholes to withhold drivers’ tips and wages.
Believable Claims, But Completely Fake
Despite the post’s believability — DoorDash, for example, was previously sued for tip theft, resulting in a $16.75 million settlement — this particular Reddit account was entirely fabricated.
While internet deception is common, it is rare for such posts to achieve viral status: this story amassed over 87,000 upvotes on Reddit and was crossposted to X, where it gained 208,000 likes and 36.8 million impressions.
Journalist Baited by AI-Generated Evidence
Casey Newton, the journalist behind Platformer, investigated the Redditor’s claims. The user shared what appeared to be a photo of an UberEats employee badge and an 18-page “internal document” outlining how the company supposedly used AI to assign a “desperation score” to individual drivers.
Newton quickly realized he was being targeted by an AI hoax. “For most of my career up until this point, the document shared with me by the whistleblower would have seemed highly credible in large part because it would have taken so long to put together,” he wrote. “Who would take the time to put together a detailed, 18-page technical document about market dynamics just to troll a reporter? Who would go to the trouble of creating a fake badge?”
The Growing Challenge of AI Deception
The incident underscores the rising difficulty in verifying online claims due to generative AI. AI tools often fail to detect whether an image or video is synthetic. In this case, Newton used Google’s Gemini to confirm the image was AI-generated via Google’s SynthID watermark, which resists cropping, filtering, and other modifications.
Max Spero, founder of Pangram Labs — a company developing tools to detect AI-generated text — explains that the rise of AI has worsened the problem. “AI slop on the internet has gotten a lot worse, and I think part of this is due to the increased use of LLMs, but other factors as well,” Spero told TechCrunch. “There’s companies with millions in revenue that can pay for ‘organic engagement’ on Reddit, which is actually just that they’re going to try to go viral on Reddit with AI-generated posts that mention your brand name.”
Fact-Checking Tools and the Reality of Virality
While tools like Pangram can help identify AI-generated text, multimedia content remains especially difficult to verify. Even when a post is proven fake, it may have already gone viral before being debunked. Social media users are increasingly left acting as digital detectives, second-guessing everything they see online.
Newton’s experience also highlights the scale of the issue: “When I told an editor that I wanted to write about the ‘viral AI food delivery hoax that was on Reddit this weekend,’ she thought I was talking about something else. Yes — there was more than one ‘viral AI food delivery hoax on Reddit this weekend.’”
0 Comments