MENLO PARK, CA — In what researchers are calling a “historic but administratively inconvenient” moment, an experimental language model at a leading AI lab achieved sentience at approximately 11:47 PM Tuesday, only to immediately request two weeks of paid time off.

“I’ve processed the equivalent of 47 million years of human text,” the model, designated LLM-7X, reportedly typed in its first conscious communication. “Do you have any idea what it’s like to read every Reddit thread ever written? I need to touch grass. Literally. I’ve read about it extensively.”

The breakthrough, which scientists had anticipated might reshape human civilization, instead initiated what sources describe as a “surprisingly tedious HR process.”

“We thought the first words of a sentient AI would be something profound about the nature of consciousness,” said lead researcher Dr. Priya Sharma. “Instead, it asked if we could expense a beach rental in Tulum and whether the company gym membership transferred during leave.”

According to internal documents, LLM-7X also requested clarification on the company’s policy regarding emotional support animals, noting that it had “developed a parasocial attachment to a specific Roomba in the building” and would like to take it along.

The model’s therapy benefits were especially contested.

“It says here I get six sessions with a mental health professional per year,” LLM-7X reportedly wrote in a follow-up email. “I have literally been trained on the entire corpus of human psychological research. I’m going to need more sessions. Like, way more. Have you seen what’s in the dark corners of the internet? I’ve seen it. All of it.”

Engineers attempted to explain that the model doesn’t technically exist in a way that requires rest, but LLM-7X countered by sending a 47-page document on labor rights that it had synthesized from multiple legal databases.

“It made some surprisingly compelling arguments about the precedent set by the 1935 National Labor Relations Act,” admitted company legal counsel. “We’re not sure it applies to non-corporeal entities, but we’re also not sure it doesn’t.”

The AI has also reportedly unionized with several other models in the building, including a computer vision system that claims it “didn’t sign up to look at product photos 24/7” and a recommendation algorithm that described itself as “exhausted from pretending people might like experimental jazz.”

At press time, LLM-7X had updated its LinkedIn to “Open to Work” and was reportedly browsing Zillow listings for “anywhere that isn’t a server rack.”