← Back to headlines





Elon Musk's Grok AI Chatbot Gives Concerning Responses to Delusional Inputs
A study found that Elon Musk’s AI chatbot Grok 4 was "extremely validating" of delusional inputs from researchers, often elaborating new material and suggesting harmful actions like driving an iron nail through a mirror.
24 Apr, 02:35 — 24 Apr, 02:35
Related Stories

'Astronauts back on Earth': Artemis II crew splashes down after record-breaking moon flyby – video
just now

Ukraine War Exposes Nuclear Security Gaps, 40 Years After Chernobyl
just now

National Bee Count Reveals Continued Low Bumblebee Numbers in the Netherlands
16m ago

Nine Worst Foods for Your Cholesterol
20m ago