Search latest AI realm updates by topic and publish date.
AI safety is becoming core industry strategy as jailbreak risks, agent misuse, cyber-defense models, text watermarking, and disclosure pressure converge.
OpenAI's acquisition of presentation startup NextSlide signals a bigger shift: ChatGPT is moving from chat assistant toward documents, decks, and everyday workplace productivity.
OpenAI limited GPT-5.6 access after a U.S. government request, turning a frontier model release into a test of AI safety review and market access.
Adobe is acquiring Topaz Labs, according to TechCrunch, signaling that AI upscaling, restoration and enhancement are becoming core Creative Cloud workflow features.
Noam Shazeer’s move from Google Gemini to OpenAI shows how the frontier AI race is increasingly being fought over elite researchers, not just models and chips.
John M. Jumper, one of the scientists behind AlphaFold's Nobel Prize-winning protein-structure breakthrough, is reportedly leaving Google DeepMind for Anth
FERC is pushing grid operators to accelerate data-center interconnections, but AI infrastructure still faces a deeper bottleneck: not enough deliverable power.
Z.ai released GLM-5.2 on Hugging Face as an MIT-licensed open model for long-horizon coding agents, claiming a stable 1M-token context window, IndexShare sparse-attention efficiency, speculative-decoding gains, and benchmark results near closed frontier models on extended engineering tasks.
Google filed a civil lawsuit against an alleged China-based phishing-as-a-service network called Outsider Enterprise, accusing it of using Gemini and other AI tools to generate phishing pages at scale. The case ties together fake websites, millions of fraudulent URLs, Android spam-text reports, FBI disruption work and an estimated $1.9 billion in losses since 2023.
A coalition of state attorneys general is investigating OpenAI, reportedly seeking information on ads, engagement, sycophancy, health data, and protections for minors and seniors.