GPT-5.6 is now everyone's default ChatGPT, Illinois passed the toughest AI safety law in the US, and India's courts are moving on deepfakes before parliament does.
GPT-5.6 is now the default ChatGPT for everyone
OpenAI's GPT-5.6 (which comes in three variants named Sol, Terra, and Luna) reached general availability on July 9 and is now the default model for all ChatGPT users, free and paid. The key difference from earlier versions: it can split complex tasks across multiple subagents and work autonomously for longer stretches without needing you to stay in the loop.
Why it matters: If you opened ChatGPT today, you're already using a significantly more capable model than you were last week, without changing any settings. The bar for what counts as "default" just moved up again.
Illinois signs the toughest AI safety law in the US
Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act on July 6, requiring companies to publicly disclose their safety practices, report major AI incidents, and submit to independent third-party audits by qualified experts with no financial conflicts of interest. Illinois, California, and New York together account for roughly 40% of the US AI market, making this effectively a national standard, even without federal action.
Why it matters: For the first time in the US, AI safety is not self-declared. A company can't just say "we're responsible", an outside auditor with legal accountability has to verify it. This is a big shift from how AI has been regulated so far.
India's courts are moving on deepfakes before parliament does
Three significant court actions on AI deepfakes hit Indian headlines in the first week of July. The Delhi High Court ordered the removal of AI-generated deepfake posts targeting an MP. The Madras High Court directed police to urgently investigate non-consensual deepfake images of an Indian woman. And the Supreme Court set aside NCLT orders that were based on AI-generated fake legal precedents. Separately, India's IT Ministry signalled it is now considering a dedicated AI regulatory framework rather than handling AI under existing rules.
Why it matters: India's regulatory framework doesn't yet cover AI specifically, but its courts aren't waiting. Real harm is being remedied now through existing laws, while the bigger policy picture takes shape. Students and professionals: what is and isn't legal around AI-generated content in India is actively being decided.
AI fake job postings are surging, and two sectors are losing 28,000 jobs a month
Two distinct but connected workforce stories are running at the same time. First: AI-powered fake job postings are proliferating rapidly, with scammers using AI to impersonate legitimate employers at scale. Employment agencies warn that the volume and sophistication of these scams is accelerating. Second: US government payroll data shows that the financial services and information technology sectors (where AI adoption is fastest) are shedding roughly 28,000 jobs per month on average. Workers with AI skills, however, are earning wage premiums up to 56% higher than peers without them.
Why it matters: The jobs story has two speeds: displacement in sectors that adopted AI first, and premium wages for people who can use AI well. The gap between those two outcomes is AI literacy. If you're job hunting right now, also know that the process itself has become a target for AI fraud.
Global AI funding hit a record $510B in just six months
Crunchbase data released this week confirms that global startup investment reached $510 billion in the first half of 2026 alone, more than the entire previous year. AI drove the surge, with nearly 40 AI startups reaching unicorn status ($1B+ valuation) in H1. Together AI: a platform for running open-source AI models at enterprise scale, closed an $800 million Series C this week at an $8.3 billion valuation. OpenAI and Anthropic alone accounted for $217 billion, or 43% of all startup funding globally in the period.
Why it matters: Capital at this scale signals that investors believe AI infrastructure is not a bubble phase, it's the next decade's operating layer. The concentration of money in a handful of companies, though, also means the rest of the ecosystem is fighting for what's left.
Google is giving every teacher free AI tools, no budget required
Announced at ISTE 2026, Google is making Gemini in Classroom available at no cost to all educators with a Google Workspace for Education account, which covers the vast majority of schools globally. The package includes over 30 new AI tools designed to help teachers create lesson content, personalise material for different student needs, and save time on administrative work. Google is also providing free AI practice tests for ACT and GRE exams to students.
Why it matters: This is one of the largest simultaneous AI deployments in education history. If your school uses Google Workspace, your teacher already has access. If you're a student, the free test prep tools are worth trying immediately.
OpenAI built its own chip, and it changes who controls AI's future
OpenAI and Broadcom unveiled Jalapeño, OpenAI's first custom AI inference chip, this week. Designed from scratch for running large language models (not adapted from general-purpose chips), it promises substantially better performance per watt than current options. Deployment is planned for end of 2026, with a multi-generation roadmap beyond that. OpenAI has also been in early discussions with Microsoft to run Claude inference on Microsoft's Maia 200 chips via Azure, meaning even a competitor's models could run on infrastructure OpenAI helped shape. The hardware race underneath the model race is quietly deciding who controls AI's future.
UN AI governance summit wraps without binding rules + WEF warns on entry-level jobs
The UN Global Dialogue on AI Governance concluded in Geneva on July 7 after two days of discussions involving governments, tech companies, academics, and civil society, but produced no binding agreements. The World Economic Forum released a report the same week on AI and the future of entry-level work, flagging that junior roles in knowledge industries are the most exposed to AI automation in the near term. Both stories reinforce the same theme: the institutions that govern work and technology are still catching up to the technology itself.