EU deepfake fines are now real, OpenAI's Astra proved theorems that stumped mathematicians, and India's AI startups crossed a billion dollars in funding in half a year.
EU AI Act enforcement is live: unlabeled deepfakes now carry fines up to €15M
The European Commission has begun enforcing the AI Act's transparency rules. AI-generated content and deepfakes must now carry machine-readable labels so they can be detected. Platforms that fail to comply face fines of up to €15 million or 3% of global turnover, whichever is higher. The Commission specifically flagged tighter oversight on AI companies creating deepfakes and cyber threats as an enforcement priority this month.
Why it matters: This is the first major jurisdiction to move from debating AI rules to actually enforcing them with real financial consequences. Every other government (including India's) is watching this playbook.
OpenAI's Astra proves 10 mathematical theorems that computers couldn't solve before
OpenAI's Astra model produced ten formally verified, machine-checkable proofs solving open problems in mathematics and theoretical computer science. These are not approximations: the proofs are checked by independent software, the same way mathematicians verify each other's work. This is a meaningful step beyond AI being helpful in research to AI actually doing original research.
Why it matters: If AI can generate verified scientific discoveries, the pace of progress in medicine, physics, and engineering could accelerate in ways that are hard to predict. This is the version of AI that changes careers in research and academia.
ChatGPT crosses 1 billion weekly users; OpenAI slashes model pricing by 80%
ChatGPT now has approximately one billion weekly active users, making it one of the most-used software products ever built. Alongside this milestone, OpenAI cut the price of its GPT-5.6 Luna model by 80%, bringing it down to $0.20 per million input tokens. That makes enterprise-grade AI dramatically cheaper to build with and access.
Why it matters: A billion users means AI is no longer a tool for tech workers, it is general infrastructure. The price cut means small businesses and developers in India can now afford to build on the same models that power global enterprise software.
AI deepfakes are now faking job applicants, and getting through interviews
Pindrop Security found that over one-third of job applicant profiles it analyzed were entirely fabricated using AI, fake resumes, AI-generated profile photos, and deepfake video interviews that fooled HR screening. Experian's 2026 Fraud Forecast ranks this as one of the top emerging threats for employers globally, and US states have passed 169 laws since 2022 trying to keep up.
Why it matters: The same tools being promoted as productivity boosters are being used to deceive hiring systems. If you work in HR, talent, or hiring in any sector, this is no longer a hypothetical, it is happening now.
India AI startups raised $1,067M in the first half of 2026, up 33% year-on-year
Indian AI startups closed $1,067 million across 157 deals in H1 2026, a 33% jump compared to the same period last year. Sarvam AI led with a $234 million round, building language models that cover all 22 official Indian languages. MeitY also ran a roundtable with VC firms and global tech companies on improving startup financing access under the IndiaAI Mission.
Why it matters: India's AI startup ecosystem is no longer a rounding error on a global funding slide. Sarvam's bet on Indian-language models is specifically building AI for problems that Western models were never trained to solve.
Maharashtra commits ₹10,000 crore to AI, with a startup fund and 1.5 lakh jobs target
Maharashtra's Cabinet cleared its AI Policy 2026, committing ₹10,000 crore in investment with a target of 1.5 lakh jobs by 2031. The policy includes a ₹500 crore startup fund, 12 AI incubators, and a mission to train two lakh youth in AI skills. It is framed as a state-level implementation of the national IndiaAI Mission, which carries a central government outlay of ₹10,371 crore.
Why it matters: State governments are no longer waiting for national frameworks to arrive, they are writing their own. If you are building or upskilling in Maharashtra, the support structure is being put in place right now.
India's two speeds: a billion-dollar funding boom beside an IT jobs slowdown
India's AI startup ecosystem is pulling in record investment. At the same time, India's legacy IT sector is shedding value and headcount as automation reduces demand for traditional software services. Maharashtra's AI policy bets on retraining two lakh people, but industry groups are not sure the pipeline is moving fast enough. The race to watch is whether India's workforce upskilling catches up to the pace of automation before the gap becomes a generation-wide problem.
Claude Academy goes live, an AI takes over your computer, and the UK names its first AI Minister
Anthropic launched Claude Academy, an official learning hub for understanding and using Claude, useful if you use it at work. Construct Computer launched an AI that operates a computer on your behalf: it browses, fills forms, and completes tasks while you do other things. And the UK appointed Kanishka Narayan as Britain's first dedicated Minister for AI, a signal that governments are treating AI as a portfolio requiring its own minister, not just a technology footnote in other departments.