Regulation was the story of the day: Illinois signed a landmark AI safety law, Geneva's governance summit wrapped, and India signalled it needs fresh AI-specific legislation: while researchers got a free 60-tool AI workbench from Anthropic.
Anthropic launches Claude Science: 60+ research tools in one place
Anthropic released Claude Science in beta, a workbench that bundles over 60 preconfigured tools for researchers, including literature search, data analysis, experiment design, and scientific writing support. It's available now for Pro, Max, Team, and Enterprise users, with no extra charge beyond existing plans.
Why it matters: This is the first AI product designed specifically for how researchers and students actually work, not just a general chat tool with a science prompt bolted on. If you're doing any kind of research or study, this is worth trying immediately.
Together AI raises $800M: AI infrastructure is where the money is going
Together AI, which builds the infrastructure layer that lets companies run open-source AI models at scale, closed an $800M Series C at an $8.3 billion valuation. The round signals that investors are moving from betting on AI applications to betting on the pipes and compute that make AI run reliably.
Why it matters: When the money starts flowing to plumbing rather than products, it means the industry is maturing. Open-source AI infrastructure growing at this scale also keeps AI accessible, not just locked inside a few proprietary platforms.
Illinois signs the AI Safety Measures Act, large AI companies must now be transparent
Illinois Governor Pritzker signed Senate Bill 315 into law, the Artificial Intelligence Safety Measures Act. It requires AI systems generating more than $500 million in annual revenue to meet new transparency and accountability standards. Illinois, California, and New York together account for roughly 40% of the US AI market, so three state laws are quietly doing the work that Congress has not.
Why it matters: US states are no longer waiting for federal action. If you use an AI product professionally, the companies behind it are now legally required to be more upfront about how those systems work, at least in the states where most of them are built.
Geneva's AI Governance Summit ends without a binding deal, but the momentum is real
The Global Dialogue on AI Governance brought together governments, tech companies, academics, and civil society in Geneva. 140+ countries spent two days debating how to regulate AI that evolves faster than any rulebook. No binding treaty emerged, but multiple countries announced new domestic frameworks, and the UN flagged that the window for fair global governance is narrowing.
Why it matters: International AI governance is moving from conference calls to actual rooms with actual commitments. The countries that shape these norms early will have outsized influence over how AI is built and governed for the next decade.
Google gives every teacher 30+ free AI tools, no budget required
At ISTE 2026, Google announced that Gemini in Classroom is now available globally in English at no cost for all educators with Google Workspace for Education accounts. The package includes over 30 AI tools, lesson planning, content differentiation, student practice tests for ACT and GRE, and interactive learning experiences built on NotebookLM and Gems.
Why it matters: Free AI tools in the hands of every teacher globally is a meaningful shift. It removes the "we can't afford it" excuse that was keeping AI out of under-resourced classrooms. The question now is whether teachers get the training to actually use them well.
India signals it needs a new AI law, the 2000 IT Act was written before AI existed
India's Minister for Electronics and IT publicly stated the country needs a fresh AI-specific law because the existing Information Technology Act, enacted in 2000, was simply not designed for generative AI. A Bloomberg analysis this week also highlighted India's growing AI sovereignty challenge in the context of US-China rivalry, and the pressure to build domestic AI infrastructure rather than depending on foreign models.
Why it matters: India setting its own AI rules (not borrowing from the EU Act or the US approach) would be significant. A law shaped around India's scale, languages, and workforce context could become a model for other emerging economies.
100,000+ tech job cuts in 2026, and the counter-narrative nobody is saying loudly enough
More than 100,000 tech industry jobs have been cut in 2026, with many companies explicitly citing AI automation as the reason. Meta alone cut approximately 8,000 roles in a single restructuring round. At the same time, AI infrastructure companies are hiring aggressively: Together AI just raised $800M, and AI safety, AI policy, and AI education roles are proliferating. The displacement is real, but so is the creation. The people caught in between are those who have the old skills but haven't started building the new ones yet. That gap is closing fast, and the window to get ahead of it is now, not when the next layoff announcement arrives.
Grok 4.5 is out, NVIDIA releases a fast open-weight model, and AI energy use is under political scrutiny
xAI released Grok 4.5 with stronger coding and agentic capabilities. NVIDIA published Nemotron-Labs-TwoTower, an open-weight diffusion model that delivers 2.4x higher text throughput. Separately, US senators launched an investigation into Big Tech's role in rising power costs, data centres already consume over 4% of US electricity, with projections putting that at 12% by 2028.