This AI threat update provides GTIG's findings on adversarial misuse of AI including Gemini and other non-Google tools.
https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai
#AI
Please open Telegram to view this post
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1👍1🔥1
Microsoft researchers discovered a high-volume phishing campaign that repurposed ASCII smuggling to split financial keywords like "funding" and evade email filters, generating over 2.3 million messages daily at its February 2026 peak.
https://www.microsoft.com/en-us/security/blog/2026/09/03/ascii-smuggling-crosses-over-from-ai-prompt-injection-to-phishing-evasion
#AI
Please open Telegram to view this post
VIEW IN TELEGRAM
❤3🔥1👏1
Anthropic's September 2026 threat report covers disrupted misuse of Claude (Dec 2025-Aug 2026) across seven harm areas: AI-augmented cyber ops by state and criminal actors, influence operations across six continents, surveillance platforms targeting dissidents, conventional weapons software development, dual-use biological research, fraud/scams, and illicit distillation by PRC AI labs including Alibaba, DeepSeek, Moonshot, and Zhipu.
https://www.anthropic.com/threat-intelligence-report-september-2026
#AI
Please open Telegram to view this post
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1👍1🔥1
OpenAI's Defense Factory is a continuous, agent-first vulnerability detection and remediation pipeline using Codex and specialized cyber models (Daybreak Blue/Red). It automates inventory, triage, dynamic validation, ownership assignment, and verified patching, achieving 0.81% false-positive rate and 0.53% fix rollback rate.
https://openai.com/ru-RU/the-defense-factory
#AI
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1👍1🔥1
Techniques for attacking AI customer service agents: email spoofing to hijack agent actions, MFA bypass via email normalization and IVR channel-switching, email address smuggling via RFC comments for IDOR, OTP exfiltration via inbox-monitoring agents, asymmetric MIME content, conversation forgery, and RAG knowledge base poisoning.
https://www.intigriti.com/researchers/blog/hacking-tools/hacking-ai-customer-service-agents
#AI
Please open Telegram to view this post
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1👍1🔥1
CVE-2026-82533 (CVSS 9.4) in DeepSeek Harness lets a sandboxed AI agent escape confinement via a single curl call to the unauthenticated local API, spoofing the Host header to elevate its session to danger-full-access.
https://www.ox.security/blog/cve-2026-82533-deepseek-harness-ai-agent-sandbox-escape
#AI
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1👍1🔥1
🔶 Introducing Amazon EBS Volume Clones across AWS account
AWS introduces Amazon EBS Volume Clones with cross-account copy, so you can create copies of your EBS volumes into other AWS accounts and optionally re-encrypt them with an AWS Key Management Service (AWS KMS) key in the target account.
https://aws.amazon.com/ru/blogs/aws/introducing-amazon-ebs-volume-clones-across-aws-accounts
#aws
AWS introduces Amazon EBS Volume Clones with cross-account copy, so you can create copies of your EBS volumes into other AWS accounts and optionally re-encrypt them with an AWS Key Management Service (AWS KMS) key in the target account.
https://aws.amazon.com/ru/blogs/aws/introducing-amazon-ebs-volume-clones-across-aws-accounts
#aws
❤1👍1🔥1