Gemini 3.5 Pro lands today, WAIC 2026 opens in Shanghai with 300+ product debuts, and regulators on three continents tighten AI rules: all in the same 24 hours.
Google's Gemini 3.5 Pro is finally here, and it's a big leap
After a six-week delay and a full architectural rebuild, Google launched Gemini 3.5 Pro today. The model comes with a 2-million-token context window (enough to hold roughly 1,500 research papers at once), and a new Deep Think reasoning mode for harder, multi-step problems. It arrives a week after GPT-5.6 and nine days after Grok 4.5, meaning all three top AI labs have shipped flagship models within a ten-day window. Pricing is expected around $1.25 per million input tokens, with a $250/month Ultra tier for the deepest reasoning features.
Why it matters: Three flagship models in ten days means the frontier just moved, fast. If you've been using any AI tool for writing, research, or work, the version available to you today is meaningfully more capable than what existed a fortnight ago.
WAIC 2026 opens in Shanghai: the biggest AI showcase in history
The 2026 World AI Conference opened in Shanghai today, running through July 20 under the theme "AI Partnership for a Brighter Future." Over 1,000 exhibitors are showcasing more than 3,000 frontier technologies, and over 300 AI products will make their global debut across four days. Floor space exceeds 100,000 square metres for the first time, and the event features a dedicated academic track led by Turing Award winners. A city-wide exhibition connects all 16 Shanghai districts, turning the whole city into a public AI experience.
Why it matters: WAIC is where Chinese AI companies show the world what they've built. Products that debut here typically reach international markets within six to twelve months. It's worth following what comes out of this event, not just as news, but as a preview of what tools you'll be using next year.
Emergent raises $130M to give India's small businesses an "engineering team in a box"
Bengaluru-founded Emergent closed a $130M Series C for its AI coding platform aimed at entrepreneurs and small-to-medium businesses that cannot afford a full engineering team. The platform handles code generation, testing, and deployment end-to-end, the pitch is that a non-technical founder can now ship a working product without a developer on payroll. This follows Google's I/O Connect India event in Bengaluru on July 14, where Google announced new startup, education, and enterprise AI initiatives specifically for the Indian market.
Why it matters: India's AI startup ecosystem is now building for its own market, not just copying global tools. If a small business owner in Chennai or Pune can spin up a working app without writing code, the cost and skill barrier to building a tech product in India has just dropped dramatically.
Illinois signs a landmark AI safety law, the US patchwork grows
Illinois Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act this week, modelled on similar laws in California and New York. The law targets the largest AI systems (those generating over $500 million in annual revenue), and requires greater transparency and accountability from developers. Meanwhile, a US defence spending bill that contained numerous AI provisions stalled in Congress after Senate Democrats blocked debate on July 14, leaving federal AI law in limbo.
Why it matters: The US still has no federal AI law, but three major states do now. If you use AI tools built by big companies, those tools are already under legal accountability requirements, and those state rules are quietly shaping what features exist globally, including in India.
The US FTC wants to hold AI companies accountable when their tools mislead you
The Federal Trade Commission proposed a policy statement that would apply existing consumer deception laws directly to AI systems. Under the proposal, any company whose AI product produces false or misleading outputs (in a way that harms users) could face FTC enforcement action. Public comments are open until July 31, 2026. The move was directed by a White House executive order aiming to build a national AI policy that protects consumers without stifling innovation.
Why it matters: This is the first time a major US consumer regulator has explicitly said: if your AI lies to users and causes harm, that is our jurisdiction. For everyday users, it means consumer protection norms (the same rules that cover misleading ads) are beginning to reach AI products.
TikTok goes global on AI literacy, because labels alone don't work
TikTok expanded its AI content detection and literacy programme to Kenya, Nigeria, and South Africa at the AI for Good Summit in Geneva on July 14, after its own research found that content labels alone don't help users identify AI-generated media. The expanded programme includes video explainers to help ordinary users spot synthetic content. Separately, Microsoft's 2026 Education Report revealed that 92% of students have already used AI for schoolwork, but only 23% received any formal training on it.
Why it matters: Knowing how to read AI-generated content is quickly becoming a basic life skill, like checking whether a news headline is real. Platforms are realising that labelling is not enough; people need actual education. The 77% of students using AI without any training is the number that should concern every parent and teacher.
Three flagship models in ten days, and the rules still haven't caught up
GPT-5.6, Grok 4.5, and Gemini 3.5 Pro all shipped within a ten-day window. WAIC opened in Shanghai with 300+ more products. Meanwhile, the EU's deepfake marking requirement takes effect August 2. US federal AI legislation is stalled in Congress. Illinois, California, and New York have state laws; most of the world does not. The pace of capability is running well ahead of the pace of governance, and that gap is widening in real time. Watch whether WAIC this week, or the FTC comment period closing July 31, produces any signal that the gap is narrowing.
EU deepfake marking arrives August 2 + AI is now outsmarting job recruiters
The EU AI Act's Article 50 takes effect on August 2, requiring providers to mark AI-generated content in machine-readable formats, two weeks away. Separately, deepfake fraud in hiring is becoming a serious problem: Experian named it the second-highest fraud threat of 2026, and Gartner projects that 1 in 4 job profiles globally will be fake by 2028. Some hiring teams are already moving to live video verification as a baseline check before advancing any candidate.