Kimi K3 became the world's largest open model overnight, the EU's deepfake labeling law is five days away, and India's AI startup funding just crossed $1 billion.
Kimi K3 goes fully open: the world's largest model is now downloadable
Moonshot AI released the full weights of Kimi K3 on July 27, a 2.8-trillion-parameter Mixture-of-Experts model that handles text, images, and video with a 1-million-token context window. Anyone can now download and run it. On benchmarks it beats GPT-5.5 and sits just behind Claude Fable 5 and GPT-5.6 Sol.
Why it matters: Open weights mean researchers, students, and small companies can build on frontier-level AI without paying API fees. The gap between what you can do for free versus what you pay for just narrowed significantly.
EU's deepfake labeling law: 5 days to go
On August 2, 2026, the EU AI Act's transparency rules go live. From that date: any platform deploying AI that generates or manipulates images, audio, or video must label it as AI-made. Chatbots must tell users they are talking to an AI. Failure to comply can mean fines of up to €15 million or 3% of global turnover.
Why it matters: This is the first time deepfake labeling becomes a legal requirement rather than a best practice in a major jurisdiction. Global platforms serving EU users (which is most of them) have to comply, which means the rest of the world will feel this change too.
India's AI startups cross $1 billion in funding: $219M raised in one week
Indian AI startups crossed $1.067 billion in funding in the first half of 2026, up 33% year-on-year. In the single week ending July 9, Indian startups collectively raised $219.2 million, with Yotta Data Services leading at $150M on a $4.4 billion valuation. AI infrastructure and deep tech accounted for the majority of capital.
Why it matters: Serious capital is moving toward India's AI infrastructure layer at speed. If you are thinking about a career in AI or building an AI product in India, this ecosystem is growing faster than most people realise.
India's AI law is taking shape, risk-based, light-touch, and pro-innovation
India's emerging AI regulatory framework will sort systems by risk level. Chatbots, productivity tools, and recommendation engines sit in a low-risk band with minimal rules. AI used in healthcare, finance, and critical infrastructure carries heavier obligations. The IndiaAI Mission received ₹551 crore in the July 2026 Union Budget, with the broader MeitY allocation up 52%.
Why it matters: India is choosing a principles-first approach over the EU's heavy rulebook. For students and professionals building in this space, this is a permissive climate, for now. The window will not stay open forever.
Google pledges AI literacy training for all 6 million US educators
At ISTE 2026, Google committed to making AI literacy training available to every K-12 educator in the United States, roughly 6 million teachers. It also launched a Classroom MCP server so external platforms can plug into Google Classroom, and rolled out new teacher-led AI activities directly for students.
Why it matters: When the world's largest search and productivity company bets on educator AI training at this scale, it sets the curriculum floor for a generation. AI literacy is becoming a teaching credential, not an optional extra.
August 2 is five days away, and the internet changes
From August 2, the EU AI Act requires all AI-generated content served to EU users to carry a label. Deepfakes must be machine-readable marked. Chatbots must identify themselves. Most major platforms (social media, video, messaging) serve Europe, so practically the whole internet shifts on that date. Watch which platforms comply immediately, which drag their feet, and how enforcement actually begins. The rules are clear; the teeth are untested.
AI fake job applicants cross 1 in 3 + 88% of students use AI + DEFIANCE Act
Pindrop Security found that over one-third of job applicant profiles they analysed were entirely AI-fabricated, passing initial screening undetected. A 35-country Digital Education Council survey found 88% of students and 77% of faculty now use AI daily, though faculty readiness is falling behind. In the US, the DEFIANCE Act, which lets victims of non-consensual AI-generated intimate images sue for up to $250,000 in damages, is now active law.