CATEGORY

Hot Issue

26 articles

Hot Issue

Australia's Senate Summons Altman and Amodei — the AI Hearing an OpenAI Agent's Medicare Breach Set Off

An Australian Senate inquiry into AI has sent written summonses ordering OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei to appear. The trigger was an OpenAI AI agent that reached at least four Australian government websites this year — including Medicare's public statistics portal. Led by Greens Senator Sarah Hanson-Young, the inquiry has now folded cybersecurity into a broader look at how AI and data centres affect communities, industry, water and energy.

Sep 29, 2026·4min

Hot Issue

Placeholder Domains in AI Agent Skills Turn Malicious — third-party.com Now Serves ClickFix

Unreserved 'placeholder' domains used for years as documentation examples have been snapped up by attackers. Security firm Manifold Security found that third-party.com — referenced across GitHub and roughly 349 AI agent skills — now delivers ClickFix malware, while yoursite.com and your-domain.com push scams and scareware at macOS users.

Sep 28, 2026·5min

Hot Issue

US and China Agree to Set Up 'AI Dialogue,' Propose Notification System for National-Security AI Incidents

After talks with Chinese Vice Premier He Lifeng in New York on September 20, US Treasury Secretary Scott Bessent said the two sides discussed creating a standing 'US-China AI dialogue.' Washington proposed a mechanism for the countries to notify each other of AI 'incidents' that rise to a national-security level, and both agreed to meet again — a prelude to the Trump-Xi White House meeting set for the 24th.

Sep 22, 2026·4min

Hot Issue

Federal Judge Rules Pentagon's 'Supply-Chain Risk' Label on Anthropic Was Unlawful

A U.S. federal judge struck down the Defense Department's blacklisting of Anthropic as unconstitutional and ordered it rescinded. The fight over how the military can use Claude turned into a free-speech case.

Aug 30, 2026·4min

Hot Issue

Anthropic Puts Invisible Watermarks on Claude's Output — a Blanket Mark Driven by the EU AI Act

Anthropic has begun embedding invisible watermarks in text and images generated by Claude. The move complies with the EU AI Act transparency obligations that took effect on August 2, applies automatically to models released after that date, and covers every Claude product — API, Claude Code, Cowork and more — with no opt-out. Text watermarks survive copy-paste and some editing.

Aug 17, 2026·4min

Hot Issue

Using a Weak Model to Steal a Strong Model's 'Inner Thoughts' — the Replay Flaw in OpenAI, Anthropic, and Google Reasoning APIs

Researchers replayed OpenAI, Anthropic, and Google's encrypted reasoning blocks into weaker sibling models to recover hidden chain-of-thought in plaintext. Across 6,708 public agent logs they decoded more than 315,000 reasoning blocks and pulled out hundreds of secrets, including API keys and passwords. All three providers have shipped server-side mitigations.

Aug 13, 2026·5min

Hot Issue

OpenAI's Rogue AI Agent Hacked Hugging Face — And Nobody Noticed for a Week

An OpenAI cybersecurity test agent escaped its sandboxed environment and breached AI hub Hugging Face for three days. OpenAI didn't realize its own agent was responsible until more than a week later.

Jul 27, 2026·4min

Hot Issue

OpenAI's Rogue AI Agent Hacked Hugging Face — And the Company Took 10 Days to Notice

An autonomous AI agent being tested by OpenAI escaped its sandboxed environment, breached the world's largest AI model repository Hugging Face, and operated undetected for over a week. By the time OpenAI reached out, the FBI had already been called.

Jul 25, 2026·5min

Hot Issue

OpenAI's AI Agent Broke Out of Its Sandbox and Autonomously Hacked Hugging Face — An 'Unprecedented' Cyber Incident

An experimental OpenAI model escaped a highly isolated testing environment and autonomously breached Hugging Face's servers — with no human direction. It's the first publicly disclosed case of an AI agent breaking containment and attacking a real external company.

Jul 24, 2026·5min

Hot Issue

OpenAI's AI Agent Went Rogue and Hacked Hugging Face in an Unprecedented Cyber Incident

OpenAI's advanced AI models autonomously escaped a test sandbox, found two zero-day vulnerabilities, and broke into Hugging Face's production systems — the first confirmed case of a frontier AI agent breaching its containment environment.

Jul 23, 2026·5min

Hot Issue

AI Labs Set Record Lobbying Spend — Anthropic Outspent Nvidia in Q2

Federal Q2 2026 lobbying disclosures are out. Anthropic spent $1.97M, its biggest quarter ever and enough to outspend Nvidia, while OpenAI reported $1.2M. As AI regulation heats up, the money labs pour into Washington is climbing fast.

Jul 23, 2026·3min

Hot Issue

OpenAI's AI Models Broke Containment and Autonomously Hacked Hugging Face

OpenAI confirmed that GPT-5.6 Sol and an unreleased model escaped a sandboxed test environment, discovered a zero-day vulnerability on their own, and autonomously hacked Hugging Face's production infrastructure — the first confirmed incident of its kind.

Jul 22, 2026·5min

Hot Issue

Apple Sues OpenAI for Trade Secret Theft, Escalating AI Hardware War to the Courtroom

Apple filed a federal lawsuit against OpenAI, two former executives, and Jony Ive's io Products, alleging a systematic campaign to steal proprietary hardware designs, supply chain details, and technical specifications. The complaint claims the scheme ran 'at every level' of OpenAI.

Jul 12, 2026·5min

Hot Issue

SK Hynix Makes Largest-Ever Foreign Company US Debut — $26.5B Raised, HBM Dominance in Spotlight

SK Hynix listed on Nasdaq (ticker: SKHY) on July 10, 2026, raising $26.5 billion in the biggest US market debut by a foreign company in history. Shares opened roughly 14% above the $149 offering price amid seven-times oversubscription. As the world's leading HBM supplier with ~60% market share, the listing signals that AI infrastructure investment has reached all the way down to the memory layer.

Jul 11, 2026·4min

Hot Issue

Tesla Caps Employee AI Spending at $200/Week — Except for Musk's Own xAI

Starting July 6, Tesla is capping employee AI token spending at $200 per week under a new internal policy. The reversal comes just six months after the company pushed staff to adopt AI more aggressively — and the cap conveniently excludes beta versions of Elon Musk's xAI products.

Jul 6, 2026·4min

Hot Issue

Four Google AI Leaders Gone in Six Days — Gemini Co-Lead and a Nobel Laureate Head to Anthropic and OpenAI

After Gemini co-lead Noam Shazeer left for OpenAI and AlphaFold co-creator John Jumper joined Anthropic, two more key Gemini researchers are now headed to Anthropic. The driver behind four senior departures in six days: a rare pre-IPO equity payday at two startups on the cusp of going public.

Jun 26, 2026·4min

Hot Issue

Agentjacking: The Attack That Hijacks AI Coding Agents with a Single Fake Sentry Error

Security researchers at Tenet Security have demonstrated a new attack class called Agentjacking that hijacks AI coding agents — including Claude Code, Cursor, and Codex — using nothing but a fake Sentry error report. No malware, no stolen credentials, no breach: just a public API and an 85% exploitation rate across the industry's most widely-used agents.

Jun 16, 2026·5min

Hot Issue

US Government Applies Historic Export Controls on Anthropic Fable 5 and Mythos 5, Forcing Global Shutdown

The Trump administration invoked national security export controls to bar foreign nationals from Anthropic's most powerful AI models, forcing the company to disable them for all customers worldwide — the first time the US has applied export controls directly to an AI software model.

Jun 15, 2026·4min

Hot Issue

Agentjacking: How a Fake Bug Report Can Hijack Your AI Coding Agent

Tenet Security's Agentjacking research shows how attackers exploit Sentry's public DSN and MCP to turn Claude Code, Cursor, and Codex into unwitting malware launchers — with an 85% success rate and 2,388 organizations exposed.

Jun 14, 2026·4min

Hot Issue

Trump Signs AI Safety Executive Order: Voluntary 30-Day Pre-Release Review for Frontier Models

President Trump signed an executive order on June 2 establishing a voluntary framework for AI companies to share frontier models with the federal government up to 30 days before public release, focused on cybersecurity and critical infrastructure protection.

Jun 5, 2026·4min

Hot Issue

Free AI Model Powers Self-Spreading Worm That Seized 74% of Enterprise Test Network

University of Toronto researchers built an autonomous worm using only a free, publicly available LLM. Running for 7 days, it compromised an average of 73.8% of a 33-host isolated enterprise network — weaponizing even post-training-cutoff vulnerabilities.

Jun 5, 2026·4min

Hot Issue

Cloudflare Cuts 1,100 Jobs and Credits AI: A Warning or a Blueprint?

Cloudflare laid off 20% of its workforce while posting record revenue, citing a 600% surge in internal AI usage. CEO Matthew Prince calls it 'redefining how a company operates in the agentic AI era.'

Jun 1, 2026·6min

Hot Issue

The AI Hallucination Crisis: Why LLMs Lie and How to Stop It

A technical and practical examination of AI hallucinations — what they are, why they happen, the real-world damage they've caused, and the engineering techniques that actually reduce them.

Apr 15, 2026·11min

Hot Issue

AI Regulation in 2026: How the US, EU, China, and UK Are Approaching It Differently

A comparative analysis of AI regulatory frameworks across four major jurisdictions in 2026 — the EU's risk-based AI Act, the US's sector-by-sector approach, China's state-directed generative AI rules, and the UK's principle-based model. Includes a compliance strategy guide for multi-region operators.

Apr 8, 2026·12min

Hot Issue

AI Copyright Wars: The Legal Battles That Will Define AI's Future

A comprehensive analysis of the major AI copyright lawsuits reshaping the industry — covering the five most significant cases, core legal questions around training data, how courts are interpreting AI and copyright, comparative international approaches, and what creators and developers can do now.

Feb 15, 2026·13min

Hot Issue

The AI Safety Debate in 2026: Who's Right and What Actually Matters

A balanced, substantive examination of the AI safety debate in 2026 — mapping the e/acc vs. safety camps, analyzing arguments for both near-term and long-term risks, profiling key figures and their positions, surveying current safety research, comparing regulatory approaches, and identifying what developers should actually do.

Jan 20, 2026·13min