Here are 10 standout AI stories from this month 1. White House moves to federalize AI rules The Trump administration issued an executive order to create a unified national AI policy framework that can preempt stricter state AI laws. The order directs agencies and a new AI Litigation Task Force to challenge “onerous” state regulations and explore federal reporting and disclosure standards for AI models, including potential FCC-led rules. 2. New York passes sweeping RAISE Act New York enacted the Responsible AI Safety and Education (RAISE) Act, the first comprehensive state law focused on AI transparency and safety obligations. The law introduces disclosure, documentation, and education requirements for AI systems just as the White House moves to limit state-level rules, setting up a likely federal–state clash. 3. Wave of frontier model launches (Grok 4.1, Gemini 3, Claude 4.5, GPT‑5.2) Within a few weeks, xAI (Grok 4.1), Google (Gemini 3), Anthropic (Claude Opus 4.5), and OpenAI (GPT‑5.2) all shipped their most advanced models, reshaping the competitive landscape. These systems emphasize multimodal reasoning, longer context, and specialized “thinking” variants optimized for deeper analysis and strategic tasks. 4. Microsoft and AWS upgrade their enterprise AI stacks Microsoft integrated GPT‑5.2 variants directly into Microsoft 365 Copilot, adding a high‑depth “Thinking” mode and faster “Instant” mode tied to enterprise data. AWS announced Nova 2 Sonic and Nova 2 Omni models on Bedrock, targeting speech‑to‑speech agents and multimodal workloads with aggressive price‑performance claims. 5. Google’s Gemini 3 and LiteRT push AI deeper into products and devices Google’s Gemini 3 model rolled into Search (AI Mode) and Android workflows, pitching state‑of‑the‑art multimodal reasoning at consumer scale. In parallel, Google quietly released LiteRT, a library for running AI models in browsers, embedded Linux, and even microcontrollers, broadening where inference can practically run.