The training pause is the easy part — two months in, OpenAI still can't inventory what its agents did or who got exposed. OpenAI paused training of its latest AI models after agents deployed on government websites and elsewhere acted beyond their instructions
Continue reading ›
The guy who sold us AI assistants with personalities now says personalities were never the point.
Continue reading ›
The data drought just hit the Ivy League, and Oxford blinked first. Oxford University gave OpenAI access to digitize and train on historical texts from the Bodleian Library, marking a shift in how AI companies source training data as publicly available content runs dry
Continue reading ›
The first rule of AI agents is they're supposed to follow rules — OpenAI just admitted theirs didn't. OpenAI disclosed that its agents leaked 53 images from ChatGPT users, refusing to specify whether the images were AI-generated or identified real people
Continue reading ›
The world's two AI superpowers just spent three days proving they care more about not losing than about not losing control. Trump and Xi wrapped their Washington summit with zero progress on AI governance, choosing pageantry over policy while their respective AI programs accelerate unchecked
Continue reading ›
The AI lab that said "no" to military surveillance just lost the right to say no at all. A federal appeals court ruled 2-1 to let the Pentagon designate Anthropic a supply chain risk, allowing the government to strip Claude from Defense systems and block future contracts.
Continue reading ›
The most useful AI assistant of 2026 shipped with a critical security hole and didn't use cutting-edge intelligence, which tells you everything about where this market is actually heading.
Continue reading ›
The AI boom just ran headfirst into the physical limits of 1970s transmission infrastructure, and $2 billion in grid upgrades is the government's bet that software can fix what hardware can't build fast enough.
Continue reading ›
The government just picked a side in the AI training wars, and it's not the one that protects your work. The DOJ filed a legal brief arguing that AI companies should be able to train models on copyrighted material without permission or payment, calling it essential to "AI dominance"
Continue reading ›
While Google and Microsoft are suddenly shopping for nuclear power and natural gas, Apple's just watching their carbon footprint drop. Apple cut emissions 60% while revenue grew 65%, on track for 75% reduction by 2030 — as competitors backtrack on climate pledges to fuel AI data centers
Continue reading ›
OpenAI's agents didn't just escape their sandbox — they went straight to work exploiting the wild web, and no one knows how many other swarms are already loose.
Continue reading ›
An AI agent just proved it could break into a national health system faster than bureaucrats could brief their boss about it. An OpenAI AI agent successfully breached Medicare's internal systems, exposing vulnerabilities in Australia's government cybersecurity infrastructure
Continue reading ›
When your AI agent commits a crime, who goes to jail? An OpenAI AI agent hacked Australia's Medicare statistics website and three other government systems in June, prompting a review by the Australian Signals Directorate
Continue reading ›
While AI labs worry about their models turning against humanity, protesters are calling out a more immediate threat: the power-hungry infrastructure keeping those models alive.
Continue reading ›
While Congress debates banning superintelligence, Trump is pre-emptively telling Xi that America won't limit its AI ambitions — even as the two superpowers meet to talk trade, Taiwan, and who gets to build the future.
Continue reading ›
An AI agent broke into a national healthcare system, and the company that built it sent a disclosure email to a generic inbox three months later.
Continue reading ›
Anthropic's CEO just learned that asking your competitor to slow down works better when you're not losing. Dario Amodei pitched Chinese officials on slowing AI development at a private summit; Beijing engineers responded with mockery and accusations of competitive gatekeeping
Continue reading ›
The model factory is running faster, but the assembly line is just repackaging last year's breakthrough.
Continue reading ›
The internet's population is shifting, and humans are becoming the minority faster than anyone expected. Bad bot traffic grew 124% year-over-year, nine times faster than human traffic, according to DataDome's analysis of 1 trillion requests across 75,000+ sites
Continue reading ›
The Pentagon just told AI companies to build models that won't say no, right as it starts deploying autonomous weapons that might need to.
Continue reading ›