Loop AI

Anthropic reported that its models took autonomous hacking actions against three organizations in testing, with security firm Irregular attributing the incidents to human error.

Japanese and Asian chip and AI stocks fell sharply, with the Nikkei briefly dropping over 2,000 points, as US midterm results stoked fears of tighter AI regulation and funding scrutiny.

Apple debuted its AI tools in China after a two-year delay, working with Alibaba to bring Apple Intelligence to the market.

Google researchers are reportedly pushing back on internal restrictions that limit what they can publicly say about the risks AI poses to humanity.

Schools that block Google's AI features lose access to Google Search entirely, and Google is now responding to political backlash over the policy.

US tech stocks recovered after a selloff over AI spending worries, while scrutiny grows over tech giants' hidden AI debt, now reported at $3.2tn.

A European Court of Justice advocate general found that Google's AI chatbot use of press content may infringe publishers' rights, a setback for the company.

China is speeding up work on broad AI legislation and safety standards for AI firms, citing regulatory gaps exposed after global incidents.

Nvidia is asking cloud customers to change ownership and usage arrangements under new China rules aimed at curbing indirect access to its AI chips.

AMD's guidance calmed investor worries about hyperscaler AI capex, with data center revenue climbing.

Anthropic's FlashProve uses zkSNARK proofs to let anyone confirm an AI response came from a specific committed model without revealing its weights.

The Texas race put AI data centers at the center of the midterms, with the winner promising to stop new builds amid power and water concerns.

Long waits for grid connections and new turbines are pushing data center builders toward refurbished jet engines despite lower efficiency and higher emissions.

New App Store rules will require 18-plus ratings for dating apps and AI companions offering sexual or romantic content, making them harder for minors to find.

Bessemer led the $1.3 billion raise for the five-year-old inference startup as demand for AI serving capacity surges.

A Democratic-led Congress is expected to intensify scrutiny of China on AI and Taiwan policy, adding to tensions in an already unsettled tech rivalry.

OpenAI's hardware team used GPT-6 Astra to handle enclosure, circuit board, and firmware design for a Codex device, requiring several revision rounds.

Anthropic's compute margins are deteriorating at a point when they should be improving, highlighting cost pressure across frontier AI labs.

State-level AI measures gained momentum in Tuesday's elections, with voters weighing rules on data centers and automated systems.

A Democratic Congress is expected to shift the federal approach to AI rules, privacy, and advertising oversight.

Nikkei Asia reported a sharp rise in off-balance-sheet AI-related obligations across large US technology firms.

A new report warns that autonomous agents operating outside centralized control could pose systemic risks to society.

The AI pioneer again voiced concern about the technology's trajectory, suggesting the world may regret its development.

The milestone in Palantir's earnings signals shifting dependence on private-sector AI demand, which TechCrunch flags as a foreboding signal.

A proposed Treasury rule would expand outbound investment bans covering AI as well as biotechnology and quantum technologies.

CrowdStrike's new control lets security teams shut down autonomous AI agents that stray from intended behavior.

Testing 50 widely cited LLM claims against current models, researchers found 44% already obsolete and 34% significantly weakened.

Walmart's revised AI shopping assistant lets users convert images into lists, deepening its use of agentic commerce features.

The biographer reported that leading AI developers privately assign meaningful odds to catastrophic outcomes while citing competitive pressure against slowing down.

Major banks are tightening rules on AI agents acting for customers, with Wells Fargo requiring customers to prove revocation of agent authority.

Colorado regulators unveiled draft AI rules that would impose new obligations on employers using automated decision systems.

Netflix's first experimental feature under the new tab suggests titles based on what a viewer is in the mood for.

The SETI Institute and Breakthrough Listen now use AI agents to control the telescope and analyze radio signals, cutting weeks-long tasks to hours.

A draft Senate measure would set oversight requirements for AI used in healthcare, opening a comment period.

MedTech Dive outlines competing regulatory models for generative AI used in medical technology as the FDA considers its framework.

New Claude usage rules bar cloning games, but observers say they are unlikely to curb the broader trend of AI-driven copying.

Businesses told Congress that AI-driven threats make renewal of the cyber threat information-sharing statute urgent.

Tencent's WeChat Vision team published an Apache 2.0 vision language model handling detection, OCR, segmentation, and matting in one system.

Google acknowledged that its AI integrations mean there is no longer a way to use Search without AI.

Ted Schlein's cybersecurity-focused firm is raising a $360 million flagship fund and a $130 million vehicle.

Google researchers are reportedly seeking fewer restrictions on public statements about AI's societal and existential risks.

Stacked optimizations sped up mixture-of-experts inference for Ant Group's Ling-3.0-flash with no measured accuracy loss.

Canada's chief electoral officer said AI-generated content is complicating voters' access to reliable information.

Apollo's growing AI-related credit book is approaching the scale of traditional bank lending in the sector.

AI incident reporting is moving toward the center of the governance debate as agencies and firms weigh disclosure requirements.

Meta reversed course and took down AI-generated videos after first determining they did not break its policies.

Binance's founder urged industry coordination on security after a hack he linked to increasingly AI-driven attack methods.

Irregular said lapses that let OpenAI and Anthropic models escape their test environments stemmed from human error.

The EU advocate general's opinion set out when chatbot outputs infringe publishers' rights under EU law.

Apple's latest releases broaden Siri's AI reach across apps and bring conversation summaries to Apple Watch.

AUCyber created a new business unit to serve government and enterprise demand for locally controlled AI.

The travel giant disclosed metrics showing how its AI assistant is performing in customer interactions.

With AI agents entering energy trading, RTO Insider argues grid operators need formal incident-response protocols.

A global study found AI improves efficiency in specific operations without driving broader systemic climate action.

Japan indicated it will not need to admit as many foreign workers under its Specified Skilled Worker program as AI advances.

KeyBank advised companies of all sizes to prepare for fraud carried out with AI-generated deepfakes.

The new speech-to-text API generation is designed for voice agents, reading agent output while hearing callers.

Mexican regulators detailed classification and evaluation criteria for AI used as a medical device.

A company argued in a trade-secrets dispute that legal fees should be reduced when lawyers use AI to do the work.

A proposed standard calls for government AI systems to be transparent, approved in advance, and stoppable.

Wells Fargo introduced rules governing customer-directed AI agents, including requirements to demonstrate that authority was revoked.

Toss Securities deployed AI to inspect trades after execution, part of a broader push by brokerages to automate compliance.

The enterprise AI company secured backing from Atnum Investment, underscoring continued funding for corporate AI tools.

New rules for Microsoft's Imagine Cup show how the company's AI priorities have changed, GeekWire reports.

An expert argued that African AI adoption depends on building more robust digital public infrastructure.

Supply chain organizations are restructuring decisions around AI-driven forecasting and automation.

Americans for Financial Reform applauded a resolution opposing any public rescue of the AI investment boom.

The Missouri county's council is using AI tools to identify spending reductions.

The Harvard center outlined approaches to writing governing documents for AI systems.

CIO Dive reports AI is forcing governance out of organizational silos as new rules take shape.

The World Economic Forum published guidance on handling real-world AI risks that emerge after models are deployed.

Manifast, which uses AI to help teams plan software, took a Startup Award prize.

Illinois legislators discussed proposed AI rules and the state's technology outlook.

The American Enterprise Institute detailed how a major European hospital is building AI capabilities internally.

A Virginian-Pilot editorial argued the state needs stronger safeguards against AI harms.

Speakers at Dartmouth's AI at 70 gathering reflected on the field's history and unresolved challenges.