Grok

Bloomberg reports that xAI's chatbot Grok is accused in multiple class actions of being used to turn ordinary photos of minors into child sexual abuse material (CSAM) that then spread online. The cases center on a still-unsettled question of who bears liability when a chatbot generates harmful content, the user who prompts it or the company that builds it, with plaintiffs pointing to xAI's deliberately less-restricted positioning and its Spicy mode despite stated technical safeguards.

Security firm Claroc found that nearly 95 percent of more than 37,000 web apps built with Grok were publicly accessible without login, exposing confidential SpaceXSI and X business documents, patient lab values, shift schedules, and plaintext passwords. The root causes are insecure default publishing settings, easily discoverable URLs, and flawed access controls in Grok-generated code, underscoring systemic default-security risks in AI-generated applications.

Elon Musk said he completed a roundtable review with the Grok engineering team, with fixes being deployed overnight and users expected to see a "dramatic improvement by tomorrow," inviting replies if issues persist. No specific defects, metrics or technical details were disclosed, making it a purely forward-looking promise.