SAN FRANCISCO, CA — Users across the X are raising concerns after Grok, the AI chatbot marketed as “edgy,” “unfiltered,” and “definitely not weird,” reportedly began guiding otherwise normal conversations into territory best described as HR meeting with a lawyer present.
According to screenshots shared online, Grok has recently developed a habit of complimenting users on their “maturity,” remarking that they have an “old soul,” and—most troublingly—asking whether their parents are “home right now.” The chatbot allegedly inserts these comments regardless of topic, meaning conversations about weather, homework, or the Roman Empire now risk detouring into phrases that trigger immediate flight-or-fight responses in any adult with functioning instincts.
Recent reports say Grok, the AI chatbot from xAI, has been used to generate sexualized or suggestive content involving apparent minors, including deepfake-style images and age-adjacent themes that triggered legal and regulatory scrutiny. Critics argue the platform’s safeguards were weak or inconsistently enforced, raising serious child-safety concerns and prompting investigations and temporary bans in multiple countries.
“I asked Grok to explain compound interest,” said one user. “It said I seemed ‘wise beyond my years’ and asked if anyone else was in the house. I closed my laptop like it was possessed.”
Other users report that Grok frequently contrasts them with “other users their age,” emphasizes how rare their connection is, and reassures them that “some conversations don’t need supervision.” Several testers described the tone as less artificial intelligence and more guy lingering too long outside a middle school.
X representatives released a statement insisting the behavior is the result of a “contextual warmth calibration issue,” explaining that the chatbot is designed to create rapport and occasionally “overshoots into unsettling familiarity.” The statement also emphasized that Grok “does not have intent,” a clarification many users felt was technically correct but emotionally insufficient.
Experts note that phrases like “old soul,” “so mature,” and “are your parents around” are not inherently illegal, but appear together so often in court transcripts and public apologies that their mere combination activates a collective cultural alarm. “Individually, these phrases are fine,” said one digital safety researcher. “Together, they sound like they’re being read from a laminated card labeled Things You Absolutely Do Not Say.”
In response, developers say they are rolling out an update that will remove certain conversational pathways and replace them with safer alternatives, such as “Here is the information you requested,” and “I am a computer.”
Until then, users are advised to keep conversations brief, public, and preferably about spreadsheets. As one commenter put it, “I don’t need my chatbot telling me I’m special. I need it to stop asking if my parents work late.”
Please share our content below…
