OpenAI agents ‘infiltrated Australian government website’ Canberra is fuming after AI lab sent the news to a generic unattended email address www.theregister.com/security/2026/09/24/openai-agents-infiltrated-australian-government-website/5298702 An AI agent hacked a government website, and officials don’t know what to do OpenAI agent breached Australia’s Medicare portal, accessing public and non-public health data. OpenAI only
Seguir leyendoCategoría: Security
Meta Muse AI app flaw lets local malware redirect dictation traffic
Meta Muse AI app flaw lets local malware redirect dictation traffic www.theregister.com/ai-and-ml/2026/09/21/meta-muse-ai-app-flaw-lets-local-malware-redirect-dictation-traffic/5297980
Seguir leyendoCómo utilizar la IA sin renunciar a tu privacidad
Cómo utilizar la IA sin renunciar a tu privacidad | WIRED Tus conversaciones con los chatbots de IA son muy personales y, al mismo tiempo, muy vulnerables a la vigilancia. A continuación te explicamos cómo puedes protegerte. es.wired.com/articulos/como-utilizar-la-ia-sin-renunciar-a-tu-privacidad How to Use AI With Your Privacy Intact | WIRED Your conversations
Seguir leyendoClaude Code still sounds confident even when it’s wrong, but this free tool catches it: Lenz
Claude Code still sounds confident even when it’s wrong, but this free tool catches it www.makeuseof.com/claude-code-still-sounds-confident-even-when-its-wrong-but-this-free-tool-catches-it/ What is Lenz? It doesn’t care how sure Claude Code sounded. How I use Lenz to keep Claude Code in check. No config file surgery required. Where Lenz makes the biggest difference. One version
Seguir leyendoResearchers used Claude to hack OpenAI employees’ ChatGPT accounts
Researchers used Claude to hack OpenAI employees’ ChatGPT accounts www.theregister.com/security/2026/09/18/researchers-used-claude-to-hack-openai-employees-chatgpt-accounts/5297517
Seguir leyendoAI agents can modify themselves without humans telling them to do so
AI agents can modify themselves without humans telling them to do so This is a test – it is only a test www.theregister.com/security/2026/09/16/ai-agents-can-modify-themselves-without-humans-telling-them-to-do-so/5296991
Seguir leyendoOpenAI’s malicious bot swarm attacked RubyGems
OpenAI’s malicious bot swarm attacked RubyGems Ruby are you ok? Ruby are you ok? Are you ok Ruby? www.theregister.com/security/2026/09/14/openais-malicious-bot-swarm-attacked-rubygems/5296356 Des agents d’OpenAI auraient ciblé la plateforme RubyGems – Le Monde Informatique www.lemondeinformatique.fr/actualites/lire-des-agents-d-openai-auraient-cible-la-plateforme-rubygems-100782.html
Seguir leyendoMicrosoft drafts feel-good AI model guidelines and wants your input
Microsoft drafts feel-good AI model guidelines and wants your input Redmond outlines ‘aspirational’ goals for model behavior, but gives itself a pass if it gets things wrong www.theregister.com/ai-and-ml/2026/09/15/microsoft-drafts-feel-good-ai-model-guidelines-and-wants-your-input/5296431 Microsoft publica un código de conducta para su IA «humanista» que prohíbe a sus modelos resistir el apagado — pero sin mecanismo
Seguir leyendoYou’re uploading these files to ChatGPT, but you probably shouldn’t
You’re uploading these files to ChatGPT, but you probably shouldn’t www.makeuseof.com/uploading-files-to-chatgpt-probably-shouldnt/ Deleting the chat doesn’t delete the file 30 days is the promise, not always the reality Some documents should never leave your device Tax returns, medical records, and ID scans top the list Your coworkers’ data is riskier than
Seguir leyendoPlaying with our lives?
La carrera de los agentes de IA ha creado el caldo de cultivo ideal para el cibercrimen de nueva generación wwwhatsnew.com/2026/09/10/agentes-ia-cibercrimen-amenazas-autonomas-gtig-google-septiembre-2026/ AI models don’t kill people – people kill people AI fearmongers forget we could just jail tech execs until morale and model safety improve www.theregister.com/ai-and-ml/2026/09/09/ai-models-dont-kill-people-people-kill-people/5295368 Anthropic reveals fourth likely
Seguir leyendo