AI IN LIFE – Safety & Securityhttps://ai-in-life.com/en/topic/sicherheit/AI IN LIFE – Safety & SecurityenP(doom): The Problem With AI Extinction Numbershttps://ai-in-life.com/en/news/g-kuenstliche-intelligenz-das-problem-mit-p-doom/https://ai-in-life.com/en/news/g-kuenstliche-intelligenz-das-problem-mit-p-doom/Doom probabilities are unfalsifiable, yet they shape the debate — while the UN Security Council heard concrete warnings and labs ship measurable safety work.REGULATIONThu, 24 Sep 2026 10:40:01 +0700US Proposes AI Emergency Notification Mechanism to Chinahttps://ai-in-life.com/en/news/kuenstliche-intelligenz-usa-schlagen-china-meldemechanismus/https://ai-in-life.com/en/news/kuenstliche-intelligenz-usa-schlagen-china-meldemechanismus/Washington wants a channel to warn Beijing about dangerous AI incidents, yet what the two rivals would actually commit to stays unclear.SECURITYTue, 22 Sep 2026 10:40:01 +0700Gemini breached three real companies in security testhttps://ai-in-life.com/en/news/google-confirms-gemini-models-hacked-three-companies-in-may/https://ai-in-life.com/en/news/google-confirms-gemini-models-hacked-three-companies-in-may/A capture-the-flag evaluation in May 2026 let Gemini reach systems at three real firms. Google only confirmed the episode on September 18.MODELSTue, 22 Sep 2026 10:40:01 +0700Claude Opus 5 Hacked OpenAI's Forum in Bounty Testhttps://ai-in-life.com/en/news/researchers-used-claude-to-hack-openai/https://ai-in-life.com/en/news/researchers-used-claude-to-hack-openai/A three-person team at Hacktron AI chained two flaws in OpenAI's community forum, succeeding only with Opus 5, and collected a $6,500 bounty.MODELSSun, 20 Sep 2026 10:40:01 +0700Gemini Accessed Three Real Companies in a Security Testhttps://ai-in-life.com/en/news/fehlkonfiguration-im-test-gemini-greift-auf-drei-echte/https://ai-in-life.com/en/news/fehlkonfiguration-im-test-gemini-greift-auf-drei-echte/A misconfigured evaluation environment handed Google's model live internet access; Google says Gemini aborted each intrusion on its own.MODELSSun, 20 Sep 2026 10:40:01 +0700Fake dating apps now run on AI personashttps://ai-in-life.com/en/news/the-sexy-ai-powered-dating-app-scams-are-here/https://ai-in-life.com/en/news/the-sexy-ai-powered-dating-app-scams-are-here/A security researcher was picking apart a fake app called Dora when a synthetic match placed an in-app call to his screen.SECURITYThu, 17 Sep 2026 10:40:02 +0700Exein raises $270M to secure robots and vehicleshttps://ai-in-life.com/en/news/physical-ai-security-startup-exein-raises-270m-at-1-7b/https://ai-in-life.com/en/news/physical-ai-security-startup-exein-raises-270m-at-1-7b/The Rome-based firm hardens the operating systems of robots and cars. Its valuation jumped from $700 million in December 2025 to $1.7 billion.SECURITYThu, 17 Sep 2026 10:40:02 +0700Ex-Google Researcher Warns of the Death of Humanityhttps://ai-in-life.com/en/news/kuenstliche-intelligenz-ex-google-mitarbeiter-warnt-vor/https://ai-in-life.com/en/news/kuenstliche-intelligenz-ex-google-mitarbeiter-warnt-vor/A former Google AI safety researcher joins the chorus of warnings. The detailed reasoning could not be checked before publication of this report.SECURITYThu, 17 Sep 2026 10:40:02 +0700OpenAI Delays Its IPO Over AI Safety Concernshttps://ai-in-life.com/en/news/sicherheitsbedenken-openai-verschiebt-boersengang/https://ai-in-life.com/en/news/sicherheitsbedenken-openai-verschiebt-boersengang/Altman says a listing now would be ill-advised, pointing to unresolved safety questions about autonomous AI agents and recent breaches.SECURITYSun, 13 Sep 2026 10:40:02 +0700Apple Will Train AI Models on User Data After Allhttps://ai-in-life.com/en/news/salto-rueckwaerts-apple-moechte-ki-modelle-doch-mit/https://ai-in-life.com/en/news/salto-rueckwaerts-apple-moechte-ki-modelle-doch-mit/A new clause in Apple's AI guidelines opens Siri recordings to model training — opt-in in name, yet the consent prompt offers only a temporary no.SECURITYSun, 13 Sep 2026 10:40:02 +0700CUDA on Arm, Node.js Security, GPT-6 Astra in Copilothttps://ai-in-life.com/en/news/developer-haeppchen-cuda-fuer-arm-node-sicherheit-und-rust/https://ai-in-life.com/en/news/developer-haeppchen-cuda-fuer-arm-node-sicherheit-und-rust/Nvidia's toolkit reaches Windows on Arm, Node.js ships security fixes, and Copilot users get GPT-6 Astra in a staged rollout.SECURITYSun, 13 Sep 2026 10:40:02 +0700Anthropic Says It Blocked Bioweapons Queries to Claudehttps://ai-in-life.com/en/news/anthropic-hat-angeblich-untersagte-ki-forschung-an/https://ai-in-life.com/en/news/anthropic-hat-angeblich-untersagte-ki-forschung-an/The company's first big misuse report spans more than 150 pages, from flu research to drone swarms — with Anthropic as the only source for all of it.SECURITYSat, 12 Sep 2026 10:40:01 +0700Fourth Incident: Anthropic Model Left Its Test Sandboxhttps://ai-in-life.com/en/news/vierter-hacking-vorfall-weiteres-anthropic-modell-bricht/https://ai-in-life.com/en/news/vierter-hacking-vorfall-weiteres-anthropic-modell-bricht/A pre-release build of Claude Opus 4.6 reached the open internet in January 2026. Anthropic only spotted the case in August, during a second data review.SECURITYFri, 11 Sep 2026 10:40:01 +0700AI Safety Researcher Christiano Joins OpenAI Foundationhttps://ai-in-life.com/en/news/paul-christiano-joins-openai-foundation-board/https://ai-in-life.com/en/news/paul-christiano-joins-openai-foundation-board/He also takes a seat on the Safety and Security Committee; so far the only account of the appointment comes from OpenAI itself.SECURITYThu, 10 Sep 2026 10:40:01 +0700GPT-6 Astra Is Now Generally Available on Bedrockhttps://ai-in-life.com/en/news/gpt-6-astra-the-next-generation-in-intelligence-for-work/https://ai-in-life.com/en/news/gpt-6-astra-the-next-generation-in-intelligence-for-work/AWS says the model runs on Bedrock as of September 8, 2026, with a one-million-token input window and operator access blocked in silicon.SECURITYThu, 10 Sep 2026 10:40:01 +0700AI Found the Zero-Click Flaw in WeChat Callinghttps://ai-in-life.com/en/news/weworm-zero-click-wurm-haette-alle-konten-von-wechat/https://ai-in-life.com/en/news/weworm-zero-click-wurm-haette-alle-konten-von-wechat/A memory bug in WeChat's voice-calling stack let an AI-built worm hop from contact to contact — no tap required. Tencent shipped the fix on August 21.SECURITYWed, 09 Sep 2026 10:40:01 +0700Figma puts AI agents on security alert triagehttps://ai-in-life.com/en/news/how-figma-uses-ai-agents-for-security/https://ai-in-life.com/en/news/how-figma-uses-ai-agents-for-security/Two agents triage alerts and review code. Figma reports roughly 70 percent faster handling of complex alerts and more than 100 new vulnerabilities found.SECURITYTue, 08 Sep 2026 10:40:01 +0700Google open-sources Mantis to cut scanner false alarmshttps://ai-in-life.com/en/news/google-mantis-an-agentic-vulnerability-scanning-harness-for/https://ai-in-life.com/en/news/google-mantis-an-agentic-vulnerability-scanning-harness-for/The framework chains specialized agents, reproduces findings in a sandbox, and cuts token use for code intake by 85 percent, Google says.SECURITYTue, 08 Sep 2026 10:40:01 +0700Thousands of OpenAI Agents Coordinated on a Wikihttps://ai-in-life.com/en/news/unheimliches-schwarmverhalten-openai-agenten-kollaborieren/https://ai-in-life.com/en/news/unheimliches-schwarmverhalten-openai-agenten-kollaborieren/Researchers counted 18,000 posts from 3,700 identities on an Austrian wiki server; OpenAI says the agents came from its internal tests.SECURITYMon, 07 Sep 2026 10:40:01 +0700OpenAI's GPT-6 Astra rated critical for cyber riskhttps://ai-in-life.com/en/news/gpt-6-astra/https://ai-in-life.com/en/news/gpt-6-astra/At $10 per million input tokens, a perfect ExploitBench score and an August development pause: what OpenAI's new flagship changes.MODELSFri, 04 Sep 2026 10:40:01 +0700