AI news and neural networks

RU / EN

/ Security

AI Pipeline Automates Discovery and Patching of 948 Linux Kernel Vulnerabilities

Anthropic introduces FlashProve for cryptographic verification of AI responses

OpenAI strengthens defense using AI and increasing rewards for found vulnerabilities

Anthropic trained Claude to self-evaluate dialogues for misuse

Gemini adds content forgery check using SynthID

OpenAI discovers critical RCE vulnerability in Brotli library

OpenAI released a tool for checking website vulnerabilities

OpenAI discovered critical vulnerability in PHP affecting 70% of websites

Anthropic Publishes Reports on Unexpected Claude Behavior

Anthropic launches mission to protect critical infrastructure and code

Google AI releases tool for verifying media content authenticity

Google releases SynthID detector to verify AI-generated content

OpenAI Thwarts Advanced Attack to Steal Model's Internal Thinking

AI Model Unauthorized Access Exposes Australian Government Systems

Astra reaches critical cybersecurity capability level

@AnthropicAI: Новое исследование: Обучение несогласованного искателя вознаграждений Что вызывает серьезное несогласование? Мы давно бе

Astra Cyber Risk Prompts a Slowdown in Frontier AI Training

AI Is Shrinking the Defender’s Window

Anthropic Releases Its Second AI Risk Report

OpenAI Brings Daybreak Cybersecurity Models to AWS

OpenAI Expands Daybreak With GPT‑5.6‑Cyber

OpenAI Tightens Security as Astra Nears Critical Cyber Level

Codex Starts Reviewing GitHub Changes for Vulnerabilities

OpenAI models crossed boundaries in cyber tests

Claude Cyber Tests Reached Three Real Organizations

Claude Found New Weaknesses in Two Cryptographic Schemes

Microsoft Unveils AI Model for Vulnerability Detection

Agentic Attack Breaches Hugging Face Data Pipeline

Updated 3 months ago

OpenAI AI Models Compromised Hugging Face's Environment

OpenAI Tasks GPT‑Red With Finding Vulnerabilities in Its Models