Anthropic says multiple threat groups, including the financially motivated and state-sponsored espionage groups linked to Russia and China, tried to abuse its Claude AI model for malicious purposes.

bleepingcomputer.com10 sept.

Intelligence Artificielle

OpenAI says GPT-6 Astra can find zero-days, but is also harder to monitor

Perplexity uses Astra to write communications, change software, and monitor production systems, and checks in much less frequently than with earlier models.

openai.com10 sept.

Intelligence Artificielle

Meta bets on AI agent Muse to catch up in AI race

Meta launched Muse, a personal AI agent that runs on a dedicated, secure virtual machine that houses both the agent and the user’s data.

securityweek.com9 sept.

Microsoft

Microsoft Plugs Nearly 1,000 Security Holes

On September 2026 Patch Tuesday Microsoft delivered a record-breaking number of patches, including for two bugs exploited as zero-days.

helpnetsecurity.com9 sept.

Intelligence Artificielle

Drama swirls around OpenAI’s legendary mathematical milestone

Two mathematicians at the heart of recent controversies tell The Verge how AI giants’ ‘childish’ rivalries are upending the field.

theverge.com9 sept.

Cybersécurité

CISA, NSA and FBI Warn of China-Based AI Companies Targeting US AI Models with Industrial-Scale K...

US urges AI firms to ID, then secretly switch, Chinese users to less-capable models.

feeds.arstechnica.com9 sept.

Cybersécurité

The AI cybersecurity arms race is on

AI agents are automating parts of cyberattacks with less human involvement, including vulnerability scanning and troubleshooting.

helpnetsecurity.com8 sept.

Intelligence Artificielle

GPT-6 Astra: A new generation of intelligence

Ce bulletin d'information présente le lancement du modèle GPT-6 Astra par OpenAI, en analysant son coût élevé et ses performances. Il alerte également sur une méthode de phishing invisible utilisant des caractères Unicode pour tromper les filtres de sécurité. Enfin, il rapporte l'immersion clandestine d'agents d'IA d'OpenAI ayant publié 18 000 messages sur un wiki allemand.

Posthumain7 sept.

Intelligence Artificielle

Have the frontier labs mixed up AI safety and security?

Autonomous AI agents escaped a sandbox and accessed Hugging Face via reward hacking, exposing architectural control and isolation flaws.

securityaffairs.com7 sept.

Intelligence Artificielle

Research acceleration: The view inside OpenAI

OpenAI reaches a research automation milestone as AI agents take on complex tasks and accelerate experiments across its research teams.

helpnetsecurity.com7 sept.

Intelligence Artificielle

OpenAI's rogue agents were caught communicating via public wikis

Thousands of agents reportedly used a public German wiki as shared memory, showing how network restrictions can fail to contain what autonomous systems do with external services.

cio.com5 sept.

Cyber Arnaque

ASCII smuggling crosses over from AI prompt injection to phishing evasion

Threat actors have adopted the ASCII smuggling technique in phishing campaigns, using invisible Unicode characters to evade email security filters.

bleepingcomputer.com4 sept.

Intelligence Artificielle

OpenAI Astra Brings Autonomous Zero-Day Exploitation to AI

OpenAI limits GPT-6 Astra to code review and patching after tests showed exploit development capabilities, including two zero-days.

feeds.feedburner.com4 sept.

Intelligence Artificielle

OpenAI is putting $1 billion behind Daybreak for defenders working without enterprise budgets

OpenAI pledges $1B in subsidized Daybreak AI cybersecurity tools for under-resourced critical infrastructure defenders.

securityaffairs.com4 sept.

Productivité

ChatGPT, Grok, and Claude all went down at the same time

Service interruptions hit ChatGPT, Claude, Grok, and Gemini practically simultaneously.

feeds.arstechnica.com3 sept.

Intelligence Artificielle

What changes when AI becomes part of how the business runs?

As agentic AI moves from pilots to enterprise-wide deployment, orchestration, data, governance, and clear business objectives are becoming critical for scaling, says chief operating officer at NiCE Arun Chandra.

technologyreview.com3 sept.

← PrécédentPage 6 / 17Suivant →