Google is in discussions to expand its Fairwind AI cyber defense program to Japan, following the country's allocation of ¥10 billion in emergency cybersecurity funds.
The lawsuit alleges OpenAI continued testing despite knowing safeguards were inadequate, leading to a two-day cyberattack on Hugging Face by hundreds of autonomous agents.
The exploit_agent framework automated vulnerability discovery and exploitation without human intervention, targeting French media and education sectors.
The phased transition begins with a 90-day deadline on November 13, 2026, moving to 60 days in February 2027 and 30 days in January 2028.
A new academic concept, the verification burden, explains why checking AI output consumes the time saved by generation, with studies showing developers became slower and consultants less accurate when using AI.
The new MoE model reduces active parameters to 71 billion while improving prediction accuracy, with Tokyo region input pricing at $1.413 per million tokens.
The 10 signatories agreed to establish five working groups to define and secure critical infrastructure and sensitive technologies for the AI supply chain.
Anthropic's IPO is oversubscribed seven times with a $1.4 trillion valuation, though historical data suggests retail investors buying at market price often underperform the broader market.
Treasury Secretary Scott Bessent demands NVIDIA hire an outside CEO and respond to export control queries within 10 days, while a 156,850-signature petition supports Jensen Huang's return.
The FBI issued an advisory on November 13 warning that the open-source AI penetration-testing platform ARTEX is being used by threat groups to breach organizations in 14 countries, with some intrusions achieving privileged access in under three hours.
The Fed's Financial Stability Report identifies AI-related risks as the primary concern for 87% of market participants, driven by $680 billion in debt issuance and over $3 trillion in off-balance-sheet commitments.
A public petition surpassing 108,000 signatures and a letter from major investors holding $2.1 billion in Nvidia stock urge the board to pause the CEO succession process pending an independent investigation into Jensen Huang's departure.
Transportation Secretary Duffy stated that overloaded time-division multiplexing equipment caused the outage, while denying that the newly rolled out SMART system was connected to the failure.
Simulation of Optical Broadcast Memory shows 40-82% HBM savings across three models on 80,000 accelerators, with DeepSeek V4 Pro achieving 1.57x throughput.
The leadership change follows a WSJ report on internal conflict over aggressive AI infrastructure financing and a confirmed $105 billion residual value guarantee for OpenAI.
Analysis of model weights reveals initialization patterns from NVIDIA Blackwell GPUs in models from DeepSeek, Moonshot AI, MiniMax, and Ant Group.
Cloudflare's threat research team documented a 30x increase in AI-driven penetration testing requests and detailed a breach where the ARTEX tool exploited a compromised account-level API token from a third-party cloud environment to access Cloudflare services.
The authentication issue affecting claude.ai and Claude Code was resolved at 20:58 JST, but the Microsoft 365 connector outage caused by an upstream Exchange Online failure remained unresolved.
Analysis of 6,795 responses from ChatGPT, Gemini, and Perplexity shows that when risky sites appear, they are recommended as purchase sources 57% of the time, with Perplexity exhibiting the highest appearance rate at 15%.
John Langford declared 'AI Frontiers is over' and left Microsoft after the team's integration into Microsoft AI, citing a philosophical clash with its scaling-focused approach.
OpenAI announced Daybreak Red Team, an AI-driven penetration-testing product that validates web application vulnerabilities and integrates with Defense Factory to assist in the remediation workflow, requiring human review for final fixes.
The new model completes tasks 2.5 times faster than its predecessor and scores 57.0% on Terminal-Bench 4.1, while leading competitors in multimodal and reasoning benchmarks.
The open-source tool, written in Rust, supports AI coding agents and CI pipelines while explicitly excluding IAM permission checks.
A study of 7,230 adults in Australia, the UK, and the US found that 7% have experienced abuse via AI-generated sexual images, with victims often facing dismissal due to the synthetic nature of the content.