RedAccess researchers found thousands of AI-built web apps leaking sensitive data due to basic misconfigurations, including Lovable apps that queried Supabase directly without proper Row Level Security. The pattern highlights growing governance gaps in low‑code and citizen development.
Parloa positions its agentic, voice-first platform as a production-ready path off rigid IVR trees, combining Azure OpenAI with low-latency speech tech, governance, and deep integrations for enterprise contact centers.
Trial exhibits outline Elon Musk’s push to bring OpenAI under tighter control, redirect talent to Tesla, and keep a back-channel into the lab. Here’s what the records show and why it matters for AI governance and competition.
Musk attempts to control OpenAI: what the trial evidence shows Read Post »
Generative AI is supercharging phishing, impersonation, and fraud while burying useful discussion on underground boards in low-value noise. Vendor research, government assessments, and reporting point to a market that is getting both more capable and more cluttered.
Cybercriminals Are Complaining About AI Slop in Cybercrime Forums Read Post »
Spectrum-X pairs Spectrum-4 switches, BlueField-3 SuperNICs, and fabric software to bring InfiniBand-like behavior to Ethernet for large AI training and inference clusters, with vendor-reported efficiency gains and lower cost per port.
NVIDIA Spectrum-X: AI-Optimized Ethernet Fabric for Gigascale AI Read Post »
Apple agreed to a proposed $250 million class-action settlement over claims it misled iPhone buyers about AI-powered Siri and Apple Intelligence availability and capabilities. The deal, which Apple says involves no admission of wrongdoing, still needs court approval.
Apple Siri AI settlement: $250M deal over marketing of Siri and Apple Intelligence Read Post »
