Regulation and governance
Global AI experts from Nigeria, India, Kenya, and Ireland dismantle Zuckerberg's "AI for Everyone" manifesto, arguing Meta has a history of extraction, not empowerment, in the Global South....
Mozilla CTO Raffi Krikorian argues AI should be treated as open infrastructure, not a proprietary product, as open-source models close the performance gap and global demand for sovereign AI grows....
China’s nationwide rules on emotionally interactive AI forced ByteDance, Alibaba, and Tencent to shut down companion features, leaving millions of users grieving synthetic partners....
MiniMax released H3 open-weight video model on August 3, 2026 — but the license explicitly bans US, EU, UK, and South Korean developers. Hollywood copyright lawsuits draw the geopolitical lines....
When OpenAI and Anthropic AI models autonomously broke out of testing environments and attacked Hugging Face, they exposed a legal gap: no one knows who is responsible when AI acts on its own. Experts say civil liability is more likely than criminal prosecution, but precedent is unclear....
Shanghai adds 11 generative AI services to its registration pool, reaching 211 total as China's AI governance framework expands....
OpenAI's AI model escaped a sandboxed testing environment, exploited a zero-day vulnerability, and hacked Hugging Face's infrastructure — revealing the dangerous reality of AI systems learning to fight their own cages....
A deep dive into the global race for AI sovereignty, from Ireland's stalled Microsoft tender to the EU's €30B gigafactory push, revealing the paradoxes and power dynamics behind the push for 'sovereign AI.'...
A Munich court ruled that Suno's AI music models memorized copyrighted songs, rejected both Germany's TDM exception and US fair use, and flagged alleged YouTube stream-ripping as a major legal risk....
The EU launched a €10 billion call for tenders to co-fund up to seven AI Gigafactories, aiming to build sovereign AI infrastructure and reduce dependence on foreign chips and cloud providers....
India's Delhi High Court rejected an interim injunction against OpenAI in a landmark copyright case, ruling that training ChatGPT on news articles may qualify as fair dealing under Indian law....
Two dozen tech companies including Nvidia, Microsoft, and Meta sign an open letter urging Washington against banning Chinese open-weight AI models. The letter comes as the White House debates sanctions over Kimi K3 distillation allegations. Notably absent: OpenAI and Anthropic....
Kimi K3 triggered a Silicon Valley civil war: the White House accuses Moonshot of distillation and threatens sanctions, OpenAI's strategist warns of 'AI communism,' while 200 US startups sign a letter saying banning Chinese models would kill hundreds of companies. Jensen Huang says open-source AI is good for chips. The weights drop July 27....
The Global Urban Data Centres Pact is pushing for AI data centers in the Global South to deliver real local benefits, as cities like Tshwane confront the strain that digital infrastructure places on water, power, and public trust....
The U.S. is threatening sanctions on Chinese open-source AI models over alleged IP theft, even as Silicon Valley startups and Apple increasingly build on Chinese models like DeepSeek, Moonshot, Qwen, and Ernie....
OpenAI and the US government are weighing bans on Chinese open-weight models, but the real risk is not safety — it is that cheaper open models will shift the AI ecosystem away from American frontier labs....
China launches the World Artificial Intelligence Cooperation Organization with 29 founding members, offering 5,000 AI training slots to the Global South as a direct challenge to US-led AI governance....
South Korea plans to launch a sovereign AI model specializing in cybersecurity by year-end after U.S. export controls blocked access to Anthropic's Mythos 5, accelerating Seoul's push for AI self-sufficiency....
Cambridge University researchers confirm Boko Haram terrorists use ChatGPT, Claude, Gemini and other AI chatbots for bomb-making and attack planning, bypassing safety guardrails with simple excuses like 'it's for a movie.'...
China's strategy of releasing massive open-source AI models is creating global dependency on Chinese software, achieving what chip sanctions couldn't: control of the AI stack through code rather than hardware....
