72% of the German Gen Z population reports productivity gains from generative AI, but usage plummets among older, less‑educated, and lower‑income cohorts.
Mark Zuckerberg unveiled a manifesto for “personal superintelligence”: AI assistants deeply tied to individuals’ goals, context, and devices, backed by new Superintelligence Labs and hardware initiatives.
Bernard Marr argues that up to 70% of repetitive RPA tasks could be overtaken by intelligent AI agents, which learn, adapt, and handle unstructured data using LLMs and contextual tools.
According to Menlo Security, web traffic to AI tools jumped 50% in less than a year, causing organizations to face mounting risks from shadow AI, data leakage, and AI-powered phishing.
5.85B images in the LAION‑5B dataset highlight why AI Bills of Materials (AIBOMs) are “the missing link in AI risk governance, tracking model lineage, data origin, and third-party code.”
Researchers found that AI models can secretly pass traits through filtered training data, with one experiment finding a model trained only on number sequences still adopted dangerous views.
