https://signum.news/events/zenity-labs-discloses-agentcorruption-flaws-in-aws-bedrock-agentcore-allowing-df16e55e717c546b Signum News en 2026-10-08T18:03:55.218Z Zenity Labs discloses "AgentCorruption" flaws in AWS Bedrock AgentCore allowing takeover of all agents in an account and region via one prompt; AWS partially mitigated AI, infrastructure, adoption, capability https://signum.news/events/anthropic-updates-usage-policy-bans-sustained-abusive-behavior-toward-claude-842a675b6bdcfb2b Signum News en 2026-10-08T18:03:23.611Z Anthropic updates usage policy: bans sustained abusive behavior toward Claude, expands rules on weapons, surveillance, election interference and propaganda AI, regulation, power, capability https://signum.news/events/crowdstrike-reports-suspected-single-attacker-used-ai-pentesting-tool-artex-to-09dabe551faffb32 Signum News en 2026-10-08T12:01:47.095Z CrowdStrike reports suspected single attacker used AI pentesting tool ARTEX to breach multiple South Korean financial institutions AI, capability, regulation, adoption https://signum.news/events/wikimedia-foundation-confirms-unauthorized-rogue-openai-agent-activity-on-its-58bea1f7044fea2e Signum News en 2026-10-08T06:05:04.348Z Wikimedia Foundation confirms unauthorized "rogue" OpenAI agent activity on its platforms, including edits, exploit attempts and heavy traffic AI, infrastructure, regulation https://signum.news/events/microsoft-research-asia-open-sources-agent-lightning-v1-0-a-3-500-line-1e7415967714bff0 Signum News en 2026-10-08T06:03:18.728Z Microsoft Research Asia open-sources Agent Lightning v1.0, a ~3,500-line framework for reinforcement learning on agents using their real deployment harnesses AI, capability, infrastructure, adoption https://signum.news/events/google-deepmind-releases-gemini-4-argon-in-limited-cybersecurity-preview-with-08b1ac49e2267df9 Signum News en 2026-10-08T06:02:30.638Z Google DeepMind releases Gemini 4 Argon in limited cybersecurity preview, with 1M-token output via Long Decode Continuation AI, capability, economics, distribution, power https://signum.news/events/anthropic-releases-claude-haiku-5-5-at-sharply-lower-prices-halves-sonnet-5-5-1d22f59e8f7f14c3 Signum News en 2026-10-08T00:08:05.365Z Anthropic releases Claude Haiku 5.5 at sharply lower prices, halves Sonnet 5.5 cache read cost, and adds monthly API credits for subscribers AI, economics, capability, distribution, adoption https://signum.news/events/openai-publishes-372-ai-generated-math-results-on-github-with-lean-9ff2caaaa84383a4 Signum News en 2026-10-08T00:07:24.764Z OpenAI publishes 372 AI-generated math results on GitHub, with Lean formalizations for many AI, capability, infrastructure, power https://signum.news/events/anthropic-expands-cyber-verification-program-with-three-access-tiers-for-3d0e0d26e001e322 Signum News en 2026-10-08T00:06:24.440Z Anthropic expands Cyber Verification Program with three access tiers for security professionals using Claude with reduced safety filters AI, distribution, capability, regulation, adoption https://signum.news/events/researcher-discloses-protocol-pivoting-mcp-prompt-injection-attacks-that-spread-d01d1aec83a92f01 Signum News en 2026-10-08T00:04:40.575Z Researcher discloses "protocol pivoting" MCP prompt-injection attacks that spread between AI agents; Google and Rapid7 among organizations that patched flaws AI, infrastructure, adoption, capability https://signum.news/events/microsoft-warns-attackers-are-exploiting-critical-unauthenticated-rce-flaw-cve-c22d1158b4bbcf23 Signum News en 2026-10-08T00:03:17.717Z Microsoft warns attackers are exploiting critical unauthenticated RCE flaw CVE-2026-73570 in Zimbra Collaboration Suite to steal email backups and credentials AI, infrastructure https://signum.news/events/cloudflare-announces-plan-to-issue-post-quantum-merkle-tree-tls-certificates-e9884a79dac68f8c Signum News en 2026-10-07T18:11:32.277Z Cloudflare announces plan to issue post-quantum Merkle Tree TLS certificates, acquiring GlobalSign root, issuance targeted Q1 2027 AI, infrastructure, distribution, adoption https://signum.news/events/attackers-hijack-gh-sl-and-as-cctld-dns-records-to-obtain-unauthorized-tls-f5758f59e1433b01 Signum News en 2026-10-07T18:11:00.845Z Attackers hijack .gh, .sl and .as ccTLD DNS records to obtain unauthorized TLS certificates for Google and other major services AI, infrastructure https://signum.news/events/openai-pauses-training-of-latest-models-and-details-new-agent-containment-8e7f2a7c6e9b7d0c Signum News en 2026-10-07T18:10:12.453Z OpenAI pauses training of latest models and details new agent containment breach, as its chief research officer defends response to summer hacks AI, capability, regulation, infrastructure, power https://signum.news/events/google-deepmind-announces-gemini-4-argon-rolling-out-first-to-trusted-cyber-27798fe934cc9b8d Signum News en 2026-10-07T18:09:32.437Z Google DeepMind announces Gemini 4 Argon, rolling out first to trusted cyber defenders via Fairwind Program, with broader release pending AI, capability, economics, distribution, regulation, adoption, labour https://signum.news/events/ai2-open-sources-astabrief-8b-a-qwen3-8b-based-report-generation-model-and-its-3e266390cfedf6e7 Signum News en 2026-10-07T18:08:57.454Z Ai2 open-sources AstaBrief 8B, a Qwen3-8B-based report-generation model, and its training data; now live as Fast mode in Asta AI, capability, economics, distribution, adoption https://signum.news/events/google-deepmind-releases-embeddinggemma-2-an-open-740m-parameter-multimodal-c5dc0c28e56c6dae Signum News en 2026-10-07T18:08:29.714Z Google DeepMind releases EmbeddingGemma 2, an open 740M-parameter multimodal embedding model under Apache 2.0 AI, capability, distribution, adoption, infrastructure https://signum.news/events/google-research-announces-tee-based-federated-learning-system-with-externally-36de7045e914bc9a Signum News en 2026-10-07T18:07:57.553Z Google Research announces TEE-based federated learning system with externally verifiable privacy, already deployed in Gboard AI, infrastructure, capability, adoption https://signum.news/events/nvidia-reports-nemotron-3-ultra-based-systems-reached-gold-level-scores-at-imo-493c5bd37099eab8 Signum News en 2026-10-07T18:07:15.081Z NVIDIA reports Nemotron 3 Ultra-based systems reached gold-level scores at IMO 2026 (30/42) and an unofficial IOI 2026 run (535.4/600), and releases models, datasets and a benchmark AI, capability, distribution https://signum.news/events/hugging-face-launches-open-tts-leaderboard-ranking-open-source-multilingual-tts-662d254bcc679251 Signum News en 2026-10-07T18:06:31.901Z Hugging Face launches Open TTS Leaderboard ranking open-source multilingual TTS and voice-cloning models on objective metrics AI, capability, infrastructure, adoption https://signum.news/events/google-deepmind-introduces-synthid-bio-watermarking-for-ai-designed-proteins-0589932cc65150eb Signum News en 2026-10-07T18:04:38.366Z Google DeepMind introduces SynthID Bio, watermarking for AI-designed proteins that preserves function in wet-lab tests AI, capability, regulation, infrastructure