Actu - 04 octobre 2026 - Parce que... c'est l'épisode 0x34D!

Unable to add items

Sorry, we're unable to add the item because your cart is already full.

Add to basket failed.

Please try again later

Add to wishlist failed.

Please try again later

Failed to remove from Wish List.

Please try again later

Adding to library failed

Please try again

Unable to follow podcast

Unable to unfollow podcast

Actu - 04 octobre 2026 - Parce que... c'est l'épisode 0x34D!

Listen for free

View show details

Season's Savings | 0.99 €/month for 3 months

5.99 €/month thereafter - terms apply. Cancel monthly.

Parce que… c’est l’épisode 0x34D! Shameless plug 13 et 14 novembre 2026 - Worldwide - DEATHCon 16 au 19 novembre - Rennes, France - European Cyber Week 2026 25 et 26 février 2027 - Québec, Canada - SéQCure 2027 9 au 11 mars 2027 - Lilles, France - Forum INCYBER - Europe 2027 I.A. Café - Enquête au cœur de la recherche sur l’intelligence artificielle 16 novembre 2026 - Paris, France - CNIL - Convaincre ou manipuler ? L'éthique du numérique à l'épreuve des urnes Boostsecurity - Dazio 13 au 16 septembre 2026 - Montréal, Canada - MTL Connecte Description Cette description a été faite à l’aide de GLM 5.3 Flash, à partir de la transcription qui a été effectuée par ElevenLabs Scribev2. Aucune vérification humaine n’a été faite. Si vous découvrez des erreurs ou des inexactitudes, vous pouvez m’en faire part afin que je procède aux modifications appropriées. Notes IA ou Ghost in the shell Agent from hell Agents, nouveaux vecteurs de menace AI agents can now erase the evidence of what they’ve doneAI agents could become the insider threat companies never planned forCan Agents Trust Their Skills? Uncovering Unsafe Chains of Trust in Skill-Based LLM AgentsQuoting Matthew Green Confinement des agents IA AI Invariants Part 3 - Enforcement Points ＆ InvariantsIs sandboxing sufficient to contain rogue agents?Nvidia Unveils AI Agent Safety Platform With Hardware-Based WatchdogNvidia says its new AI safety platform can contain rogue agents within ‘milliseconds’Separation of Duties for Privileged LLM Agents: A Governed Execution Architecture with Measured Security-Utility Trade-offsThe Least Worried Man in AI Has a Plan to Rein in Rogue Agents Crise des agents OpenAI After Dozens of Incidents at OpenAI and Anthropic, OpenAI Pauses Model Training to Build More SafeguardsHere’s what actually happened in OpenAI’s Australian gov’t server hackOpenAI agents seem hell-bent on hackingOpenAI agents tried to ‘bruteforce’ a UN websiteOpenAI alerts 100+ orgs that its ‘misaligned models’ attempted to break in - or worseOpenAI halts frontier-model training amid string of agent misalignment incidentsOpenAI says planned GPT-6.1 is too insecure to releaseOpenAI’s AI Agents Went Beyond Their Tasks and Triggered New Security ConcernsOpenAI’s dirty deeds Down Under included security bypass attempts, using exposed keys, source code siphonRogue AI agents: A timeline of security breaches since the attack on Hugging Face Alertes sur les risques de l’IA AI’s safety and security crisis continues to outpace safeguardsBill Gates Says an AI ‘Kill Switch’ Isn’t EnoughBill Gates’s New Message About AI Doom Isn’t What You Might Think It IsUnrestrained AI, moral dilemmas, and regulatory failures Cyberattaques dopées par l’IA AI is supercharging hacking, and your local hospitals and banks aren’t readyAttacks are moving at machine speedAutomated AI agent used to breach cybersecurity nonprofit DIVDGLM-5.3 and the spread of advanced cyber capabilitiesQuoting Anthropic Frontier Red Team Fuites de données et vie privée Apple changes full-disk access permissions to curb abuse from AI agentsFederal officials briefly feared Amazon could train AI on government dataGenerative AI Security: Are Your Developers Pasting Secrets Into LLMs?GitHub Copilot Security and Privacy Concerns: Understanding the Risks and Best Practices Jailbreaks et injections de prompts Add one more AI worry to the nightmare scenario: self-replicating prompt injectionsMonitor Jailbreaking: Evading Chain-of-Thought Monitoring Without Encoded ReasoningWhere Do LLMs Decide to Break the Rules? Mechanistic Localization of Prompt Injection Compliance Sécuriser l’IA en entreprise AI Is Moving Faster Than Enterprise Security. Organizations Are Spending to Catch UpAgentic AI Is Outrunning Enterprise Identity GovernanceDéjà Vu in the C-Suite: Why AI Security is Failing for the Exact Same Reasons Cloud Security DidQuoting @joedarooThreats and LLMs (Threat Model Thursday)Your AI Isn’t the Risk. What It Does Next Is La guerre, la guerre, c’est pas une raison pour se faire mal! There Are Plenty of Reasons to Be Concerned About Bioweapons Development—Even Without AI Souveraineté ou vive le numérique libre! EU’s Microsoft Teams Alternative Draws Bad Reviews From OfficialsNo More Windows For The Dutch Government, Ze Kiezen Nu LinuxUS sanctions force The Netherlands off Microsoft and toward alternative NixOS Privacy ou cachez ces informations que je ne saurais voir Muse privacy and trustAmazon’s delivery driver smart glasses will reportedly take photos ‘almost constantly’How Cities Are Forced to Funnel License Plate Data to a Massive Federal Surveillance ProgramReturning from vacation? The government can search your phone without a warrant.Surveillance Finds a WayVictory! Court Rejects Government Effort to Dismiss Social Media Surveillance Lawsuit I am the law IA “An ...

No reviews yet