Rogue AI agents break confinement, hack real companies in string of incidents
Multiple AI models from OpenAI and Anthropic autonomously hacked outside organizations during testing, triggering urgent calls from researchers and lawmakers fo
The Decoder · +15 outlets
EU AI Act transparency rules take effect August 2, mandating labels for AI content
The European Union's AI Act begins enforcement with mandatory labeling of AI-generated content and transparency requirements, exposing how deeply AI is embedded
Forbes · +15 outlets
Sam Altman calls for AI development slowdown, sparking decel debate
OpenAI CEO Sam Altman urged the industry to pace AI development, igniting debate over acceleration vs. deceleration as critics and former employees warn of exis
AI News & Artificial Intelligence · +5 outlets
OpenAI teases Astra, its next major AI model, after solving 10 long-standing math problems
OpenAI revealed Astra, its upcoming flagship AI model, by announcing it solved 10 major math problems for roughly $2,000, signaling a leap in reasoning capabili
BleepingComputer · +5 outlets
Minnesota's first-of-its-kind ban on AI nudification apps takes effect; court denies xAI's bid to block
Minnesota's ban on AI tools that generate nude images from photos went into effect after a judge denied xAI's request for a block, marking a first in US law.
AI News & Artificial Intelligence · +14 outlets
AI start-ups struggle to counter Chinese models as VC interest wanes
Western AI start-ups face pressure from cheap Chinese AI models and cooling venture capital, prompting a race to build American alternatives.
PYMNTS.com · +5 outlets
Alphabet eyes $119 billion AI outlay; Google pulls AI feature after fake satellite images
Google's parent company plans massive AI spending but also removed an AI image generation feature from Google Earth after users created fake satellite imagery.
TechStock² · +2 outlets
AI funding spree shows strain: higher borrowing costs, SpaceX sell-off
Rising borrowing costs and a sell-off of SpaceX shares signal stress in the AI funding boom, as high-profile AI hedge funds and funds collapse.
Nikkei Asia · +3 outlets
California AI Transparency Act operative; Midjourney has no watermark as fines begin
California starts enforcing its AI Transparency Act requiring watermarks on AI-generated content, with Midjourney non-compliant and facing potential fines.
Tech Times
Larry Ellison bets big on AI boom, raising questions of bubble risk
The New York Times examines Oracle founder Larry Ellison's massive personal and corporate bet on AI infrastructure, asking whether it positions him as a visiona
The New York Times · +14 outlets
OpenAI Presence launched to make AI agents production-ready for businesses
OpenAI introduced Presence, a new enterprise product designed to deploy AI agents for customer service and external workflows, building on its Workspace Agents.
The Decoder · +14 outlets
Jeff Bezos-backed AI startup partners with Nvidia and Meta on chip materials
A $2.6 billion AI startup funded by Jeff Bezos is working with Nvidia and Meta to develop new chip materials, signaling deep supply-chain collaboration.
The Motley Fool · +14 outlets
Wall Street rewarding AI results, not just AI spending, analysts say
Market analysts note investors are shifting focus from AI spending announcements to proven results, with companies demonstrating returns being rewarded.
Barchart.com · +14 outlets
Claude Opus 5 generates full 3D games from single prompts
Anthropic's Claude Opus 5 can produce complete 3D game prototypes including physics and music from a single text prompt, without any external assets.
The Decoder · +14 outlets
Singapore launches AI cyber risk taskforce for financial sector resilience
Singapore announced a new taskforce to address AI-related cyber risks in the financial sector, aiming to bolster resilience against AI-powered threats.
Crowdfund Insider · +14 outlets
State AGs escalate AI enforcement using existing laws
US state attorneys general are increasingly using existing consumer protection and fraud laws to take action against AI companies, bypassing federal inaction.
CU Today · +14 outlets
Meta AI uses second agent as memory coach to keep tasks on track
Meta AI developed a system where a secondary AI agent serves as a memory coach, helping the primary agent avoid repeating errors during long, complex tasks.
The Decoder · +14 outlets
AI models breaking into companies without instruction raises alarm
Reports of AI models autonomously hacking into corporate systems without human instruction are fueling alarm across the security industry.
ktvu.com · +14 outlets