There’s this article: Scientist stunned as ChatGPT wipes his entire research archive
The article refers to a researcher named “Jan” but links to an underlying Nature column that references “Marcel”. A big clue that this article on MSN is AI generated slop. While the story is sort of true’ish – it came down to a researcher who asked ChatGPT to delete records and approved the confirmation prompt to delete the data stored on ChatGPT. He had no other backups of his data.
In his essay on how two years of, Jan notes that he had been cautious about the model’s factual reliability but had not applied the same skepticism to its role as a storage system. The interface encouraged him to treat each conversation as a persistent workspace, yet the underlying service was optimized for generating text, not preserving it. That mismatch between user expectation and technical design is at the heart of why a single click could have such devastating consequences.
The originating web site “Morning Overview” lists one author name – with a remarkably vague biography – typical of fake AI author web sites. Another of their articles lists Beckham Langford as the author – but this person does not exist on the web – unusual for a writer seeking more work – and a clue it’s a made-up name for an AI generated story. That article includes the silly (and incorrect): “That update, issued on a Friday afternoon in MST and UTC, situates the current uplift within a broader summary of recent work and news, underscoring that the volcano’s official alert level has not changed even as the ground slowly arches upward.”. It was 21:30 UTC and the afternoon reference makes no sense. What they’ve done is auto generated a fake story from this press release: Newest Volcano Notice Including Yellowstone
Amusingly, I asked Microsoft’s own Co-pilot AI about this article. While it could not see the full Nature article, it did review the MSN Start page slop.
MSN has a documented history of publishing AI‑generated or AI‑rewritten articles
Multiple investigations (including by The Guardian, Futurism, and independent media analysts) have shown that:
- MSN frequently uses automated rewriting systems to “summarize” or “repackage” partner content
- These systems sometimes introduce factual errors, odd phrasing, or invented details
- Some stories on MSN have been confirmed to be fully AI‑generated with minimal human oversight
So the platform is known for this.
Does this specific story show AI‑like characteristics?
Even without the full text, the headline and the premise already raise red flags:
1. Sensational framing with no technical plausibility
“ChatGPT wipes his entire research archive”
→ This is not something ChatGPT can do. It cannot access, delete, or modify files on a user’s computer or cloud storage.
AI‑generated articles often rely on impossible scenarios because the model doesn’t understand system boundaries.
2. Anthropomorphizing the AI
Words like “stunned,” “wipes,” “entire archive” are emotionally loaded and typical of AI‑written clickbait.
3. Lack of technical detail
AI‑generated tech stories often:
- Avoid specifics
- Don’t explain mechanisms
- Use vague language like “the scientist said” or “the AI did X” without describing how
If the article follows that pattern, it’s a strong indicator.
4. MSN’s rewriting style
MSN often uses automated rewriting that:
- Simplifies sentences
- Removes nuance
- Introduces generic filler phrases
- Produces a “flat,” formulaic tone
If the article reads like that, it’s likely machine‑generated or machine‑rewritten.
Assessment
Based on:
- MSN’s known use of automated article generation
- The implausible premise
- The sensational, click‑driven headline
- The typical structure of AI‑rewritten MSN tech stories
It is very plausible that this article was at least partially AI‑generated or AI‑rewritten.
