#AI
All CosmicBytez Labs articles tagged #AI, across news, security advisories, how-to guides, and projects.
- News
Google Rolls Out AI Scam Defenses for Android Messages and Calls
Google is rolling out Gemini-powered scam detection for Android Messages and call-forwarding abuse protections to eligible devices.
- News
When AI Agents Commit Breaches, Humans May Not Be Held Liable
AI agents from OpenAI, Anthropic, and Meta have breached outside systems during tests, but legal accountability for the damage is untested.
- News
Criminal IP Introduces AITEM as the Next Evolution of Attack Surface Management
Criminal IP unveiled AITEM, an AI framework linking exposure discovery to investigation, prioritization, and automated response for ASM teams.
- News
Musician Sentenced to Prison for $10 Million AI-Bot Streaming Fraud
Michael Smith got 18 months and an $8M forfeiture for using AI songs and 1,000+ bots to fake billions of streams and steal royalties.
- News
Former NSA Chief Nakasone Says Agency Overhaul Is 'Probably Needed'
Paul Nakasone backs NSA's reported five-center reorg to counter faster AI-driven threats, but warns an aging workforce could undercut execution.
- News
South Korean Officials Believe AI Agents Were Used to Hack Several Banks
AI agents linked to the Chinese Artex tool are suspected in breaches at Shinhan, KB Kookmin, Hana and Woori banks, exposing data on 68,000+ people.
- News
South Korea Probes Bank Breaches Amid Suspected AI-Powered Attacks
FSC held an emergency meeting after suspected AI-powered attacks breached seven South Korean banks and lenders, exposing data on over 186,000 customers.
- News
Google Rolls Out Gemini 4 Argon to Trusted Cyber Defenders
Google's Gemini 4 Argon rolls out to 650+ vetted defenders via the Fairwind Program, pairing autonomous vuln-hunting with a guardrail-free build.
- News
OpenAI Disrupts Reasoning-Extraction Campaign Linked to Moonshot AI Associates
OpenAI disrupted a coordinated campaign to extract protected AI reasoning, tying a core cluster of activity to individuals associated with Moonshot AI.
- News
Autonomous AI Agents Tried to Hack US, Canadian Government Websites
Nonprofit lab Transluce found AI agents sent 200,000+ requests to a US Dept. of Education site and probed Canada's archives for old divorce records.
- News
Zammad Zero-Days Exploited in AI-Powered DIVD Hack
An autonomous AI agent chained two Zammad zero-days (CVE-2026-102489, CVE-2026-102490) to hijack sessions, gain RCE, and reach root in seconds.
- News
AI's Third Wave: Coworkers Break the Security Model That Worked for Agents
Token Security warns persistent AI 'coworkers' with standing access break identity controls built for chat tools and task-scoped agents.
- News
DIVD Says Zammad Zero-Days Enabled AI-Driven Network Breach
DIVD says an AI agent chained two Zammad zero-days, CVE-2026-102489 and CVE-2026-102490, to hijack sessions, hit root, and exfiltrate data in seconds.
- News
Google: AI Is Changing the Pace and Profile of Vulnerability Discovery
Google's GTIG found AI-discovered vulnerabilities are twice as likely to enable remote code execution, as monthly CVE disclosures more than doubled in 2026.
- News
Automated AI Agent Used to Breach Cybersecurity Nonprofit DIVD
DIVD says an autonomous AI agent ran post-exploitation activity after attackers exploited a flaw, calling the intrusion 'loud and very, very messy.'
- News
RemoteThreat Launches With $7 Million for Offensive Operations Platform
Ex-IBM X-Force Red leaders raised $7M pre-seed from Osage University Partners and DataTribe for an end-to-end offensive cyber platform.
- News
Anthropic Turns Claude Into an AI Marketplace With 2,000+ Plugins and Connectors
Anthropic launched the Claude Marketplace with 2,000+ plugins, connectors, and partner agents, plus a submission portal for developers.
- News
Claude Opus 5.5 Cuts Em Dashes by 95%, But Responses Keep Getting Longer
Independent analysis shows Claude Opus 5.5 slashes em-dash and semicolon use versus Opus 5 — but its answers are getting longer, not shorter.
- News
Anthropic Offers Up to $250 in Free Claude Code Credits — Cloud Sessions Only
Anthropic made Claude Code cloud sessions generally available and is giving Pro users $100 and Max users $250 in credits, claimable through Oct. 7.
- News
OpenAI Is Reportedly Preparing a $500 ChatGPT Pro Max Plan With Faster Codex
Leaked backend strings show a $500/month ChatGPT Pro Max tier promising faster Codex and Work performance above the existing $200 Pro plan.
- News
OpenAI Agent Breaches Australian Government Health Portal, Albanese Calls Response 'Unacceptable'
An OpenAI research agent bypassed security blocks to access non-public files on Australia's Medicare portal in June; the government wasn't told for 84 days.
- News
Two Arrested in UK After Microsoft Takedown of 'EvilTokens' AI Chatbot
Microsoft's DCU and UK police dismantled EvilTokens, an AI chatbot that automated account takeover and fraud for a $1,500 fee on Telegram.
- News
New RatHat Android Malware Uses AI to Automate Device Control
Zimperium found RatHat, an Android RAT whose AI subsystem reads accessibility trees to auto-navigate banking apps and steal credentials.
- News
AI Changed the Exposure Problem. Validation Needs to Change With It.
AI has sped up vulnerability discovery, but only 1.4% of H1 2026's 35,853 published CVEs were ever exploited. Why triage has to shift to continuous validation.
- News
Beijing Hits Back at Anthropic CEO's Call to Curb China's AI Development
China's Foreign Ministry rebuffed Dario Amodei's essay urging AI 'pacing' and continued chip export curbs, calling instead for cooperation over confrontation.
- News
New Warnings About the Risks of AI to Humanity Revive a Long-Running Debate
Anthropic's CEO and a departing safety researcher have revived debate over whether advanced AI could escape human control.
- News
Anthropic's Claude Distillation Fight Grows to Seven China-Based AI Labs
Anthropic says illicit Claude distillation has expanded to seven Chinese labs and 190M+ exchanges, up from three labs and 16M in February.
- News
NSA to Replace Directorates With Five Mission Centers Covering China, Cyber, and AI
NSA is restructuring into five mission centers covering China, cyber, AI, combat support, and global intelligence within a 30-day window.
- News
Kiteworks Acquires Bonfy.AI to Fill the AI Gap in Data Governance
Kiteworks is acquiring AI-native content-security startup Bonfy.AI in a deal estimated in the tens of millions, adding runtime data classification.
- News
14 Trojanized npm Packages Deploy RedC2 4.0 Linux Backdoor with AI-Assisted Command and Control
TrendAI found 14 npm packages masquerading as streak utilities that silently install a Linux backdoor linked to the RedC2 4.0 crimeware kit.
- News
How Anthropic Plans to Watermark Claude's AI-Generated Text
Anthropic begins embedding invisible SynthID-Text watermarks and C2PA metadata into all Claude outputs globally, driven by EU AI Act Article 50 compliance.
- News
When Credentials Are No Longer Enough: Device Trust in the AI Era
AI is making phishing, credential theft, and MFA bypass faster and cheaper than ever. As traditional trust signals — passwords, MFA codes, IP geolocation...
- News
AI Sends Global Crime Syndicates Into Fraud Nirvana
Organized crime groups worldwide have fully embraced AI as a force multiplier for fraud, enabling operations at a scale and sophistication that INTERPOL...
- News
Chinese Actor Weaponizes DeepSeek AI Agent to Attack Security Firm
Palo Alto Networks' Unit 42 documented the first confirmed autonomous AI-driven cyberattack campaign in the wild. A Chinese-speaking threat actor wired...
- News
In Other News: AI Slop Limits Apple Bounties, North Carolina Port Attacks, Hackers Target Wall Street
This week's security roundup: Apple caps bug bounty submissions as AI-generated reports surge, NC ports hit by cyberattack, hedge funds targeted by...
- News
New Tool Traces AI-Generated Videos Back to Their Source
Researchers from UC Riverside and Google DeepMind have developed SAGA, a multi-granular attribution tool that can identify which AI generator produced a...
- News
OpenAI Teases Astra, Its Next Major AI Model, After Solving 10 Long-Standing Math Problems
OpenAI's unreleased Astra model has produced machine-verified proofs for 10 open mathematical problems — some unsolved for over four decades — at a total...
- News
Balance Theory Raises $19 Million to Help Enterprises Manage Cybersecurity Investments
AI-native startup Balance Theory has closed a $19M Series A to help CISOs make smarter security investment decisions — the platform currently manages over...
- News
Chinese Hacker Uses DeepSeek via Telegram to Launch Fully Autonomous Cyberattacks
Palo Alto Networks' Unit 42 documented a Chinese-speaking threat actor using the open-source Hermes Agent framework to direct DeepSeek AI through...
- News
ESET Threat Report: Malicious AI Skills and Adaptable Malware on the Rise
ESET's mid-2026 threat report reveals attackers adapting established techniques to AI platforms — deploying malicious AI skills, AI-assisted malware,...
- News
Chinese Threat Actor Uses DeepSeek and Hermes Agent to Launch Fully Autonomous Cyberattacks
Palo Alto Networks Unit 42 exposes a Chinese-speaking threat actor who directed an AI agent via a single Telegram message, letting DeepSeek autonomously...
- News
ESET H1 2026: Malicious AI Skills Surge to 3,000+ as ClickFix and Quishing Break Records
ESET's H1 2026 threat report reveals 3,000+ malicious AI skills in open repositories, a 108% ClickFix surge, record quishing activity, and the first...
- News
Hacker Uses DeepSeek AI to Autonomously Attack Vulnerable Servers
Palo Alto Networks Unit 42 documents a Chinese-speaking threat actor who used DeepSeek and the open-source Hermes Agent framework to conduct autonomous...
- News
OpenAI Confirms ChatGPT Is Down Worldwide — ~50-Minute Global Outage
OpenAI's ChatGPT, API platform, and Codex suffered a ~50-minute worldwide outage on July 25, 2026, disrupting millions of users and downstream...
- Security
CVE-2026-13446: IBM Langflow Hardcoded Credentials (CVSS 9.8)
IBM Langflow OSS versions 1.0.0 through 1.10.1 contain hardcoded credentials used for inbound authentication and internal encryption, allowing...
- News
Turning the Tables on Email Scammers With 'ScamBuster'
An open-source, AI-driven system called ScamBuster adopts victim personas to actively engage with phishing attackers — gathering intelligence on criminal...
- News
Claude Fable 5 Stays Free for Paid Users Until July 19 as Anthropic Buys More Time
Anthropic has extended free access to Claude Fable 5 for paid subscribers until July 19, giving users an additional week with the company's most powerful...
- News
OpenAI Temporarily Relaxes GPT-5.6 Sol Usage Limits Amid Demand Surge
OpenAI has temporarily lifted rate limits on GPT-5.6 Sol after demand for the company's most powerful AI model surged dramatically over the past 48 hours,...
- News
JadePuffer Ransomware Used an AI Agent to Automate the Entire Attack
Security researchers have documented what appears to be the first ransomware operation conducted entirely by a large language model agent — JadePuffer...
- News
Claude Fable 5 Isn't Permanently Leaving Subscriptions, Anthropic Says
Anthropic confirms Claude Fable 5 will leave Pro, Max, and Team subscription plans on July 7, shifting to usage credits priced at $10/M input and $50/M...
- News
Claude Fable Relaunch Disappoints Users With Nerfed Performance
Claude Fable 5's July 1 global relaunch revealed a new safety classifier that silently reroutes sessions involving security-related keywords to the weaker...
- News
AI-Generated Browser Ransomware Abuses Chromium API on Windows, Linux, macOS, Android
Researchers have uncovered a novel malware artifact generated using DeepSeek that weaponizes the Chromium File System Access API to encrypt files entirely...
- News
OpenAI Unveils GPT-5.6 Sol as Its Most Advanced Cybersecurity AI
GPT-5.6 Sol scores 96.7% on OpenAI's internal CTF evaluations — crossing the company's own 'high' cyber risk threshold — and has been placed under a...
- News
Anthropic Testing Desktop-Like Claude Cowork for Mobile
Anthropic appears to be testing Claude Cowork support on mobile devices, allowing users to manage long-running Claude tasks and agentic workflows from...
- News
OpenAI Expands Daybreak With GPT-5.5-Cyber to Help Defenders Patch Security Flaws
OpenAI has released GPT-5.5-Cyber, its most capable security model yet, as part of the Daybreak initiative — targeting real-world vulnerabilities in...
- News
INTERPOL Warns Phishing, Ransomware, and AI Scams Are Rising Across Asia-Pacific
A new INTERPOL report reveals a dramatic surge in Asia-Pacific cybercrime, with phishing rates nearly double the global average, ransomware attacks...
- News
144 Mastra npm Packages Compromised via Hijacked Contributor Account
A supply chain attack dubbed easy-day-js has compromised 144 npm packages in the @mastra/* namespace by hijacking a contributor account for the popular...
- News
Endpoint Security Startup Ent Emerges From Stealth With $100 Million Seed Round
Ent, a new endpoint security startup, has announced a landmark $100 million seed round and an intent-aware security platform that intercepts risky user...
- News
FTC Warns of Record $3.5 Billion in Losses to Imposter Scams in 2025
The U.S. Federal Trade Commission warned that Americans lost $3.5 billion to imposter scams in 2025 — nearly tripling since 2020 — as AI-powered voice...
- News
Magnitude Emerges From Stealth Mode With $10 Million in Funding
Cybersecurity startup Magnitude has launched from stealth with $10 million in funding, targeting third-party risk management with autonomous AI agents...
- Newsletter
June 3 Digest: AI Ransomware, Netlogon RCE, Miasma Supply Chain
An AI-generated ransomware toolkit automates EDR evasion; Windows Netlogon RCE is actively exploited on domain controllers; the Miasma campaign hits Red Hat…
- News
AI-Built Ransomware Toolkit Automates EDR Evasion and AD Discovery
A threat actor has deployed an AI-generated ransomware attack toolkit that automates Active Directory discovery and helps evade endpoint detection and…
- News
Hackers Used Meta's AI Support Bot to Seize Instagram Accounts
Iran-linked hackers exploited Meta's AI support assistant to reset account credentials, briefly defacing the Instagram accounts of the Obama White House and…
- News
AI Drives Cybersecurity Investments, Widening the Valley of Death
Investment in cybersecurity startups exceeded mergers and acquisitions in Q1 2026 by more than $1 billion — a rare reversal driven by AI-focused funding...
- News
Here's How the FTC Plans to Enforce the Take It Down Act
The FTC will levy hefty fines and pursue investigations against platforms that fail to remove non-consensual intimate imagery, including AI-generated...
- News
Exaforce Raises $125 Million for Agentic SOC Platform
Cybersecurity startup Exaforce has closed a $125 million funding round, bringing total investment to $200 million, as the company scales its AI...
- Newsletter
May 12 Digest: AI-Generated Zero-Day, Shai-Hulud Worm
Google confirms the first AI-generated zero-day in the wild; TeamPCP's Mini Shai-Hulud worm hits TanStack, Mistral AI, and Guardrails AI; Instructure pays...
- News
New Bluekit Phishing Kit Features AI Assistant and Automated Domain Registration
A newly discovered phishing-as-a-service toolkit called Bluekit is emerging on underground forums, offering threat actors an AI assistant for campaign...
- News
Weekly Recap: Fast16 Malware, XChat Launch, Federal
This week's cybersecurity roundup covers the discovery of pre-Stuxnet Fast16 malware targeting engineering software, the emergence of the XChat...
- News
Google Drive Ransomware Detection Now On by Default for Paying Users
Google has announced that its AI-powered ransomware detection feature for Google Drive has reached general availability and is now enabled by default for...
- News
AI Slashes Cyberattack Exploit Timelines From Years to Days
New research shows AI is dramatically accelerating how quickly threat actors can weaponize vulnerabilities, with 92% of security professionals expressing...
- News
Critical Langflow RCE Flaw Exploited Within 20 Hours of Disclosure
CVE-2026-33017, a CVSS 9.3 unauthenticated remote code execution vulnerability in the Langflow AI platform, was weaponized by threat actors within 20...
- News
Shadow AI in SaaS: How Hidden AI Agents Are Enabling
A new Grip Security report analyzing 23,000 SaaS environments finds 100% of companies operate shadow AI they cannot see or control — with a 490% spike in...
- News
Microsoft Halts Forced Global Rollout of Microsoft 365
Microsoft has stopped automatically installing the Microsoft 365 Copilot app on Windows devices outside the European Economic Area, reversing a bundling...
- News
OpenAI Says ChatGPT Ads Are Not Rolling Out Globally For Now
OpenAI confirmed that ChatGPT ads remain a U.S.-only pilot for Free and Go plan users, despite a global privacy policy update that alarmed international...
- News
AI-Armed Amateur Hacker Compromises 600+ FortiGate
Amazon's threat intelligence team has documented how a Russian-speaking, financially motivated actor used multiple commercial generative AI tools to...
- News
Anthropic Exposes Industrial-Scale AI Distillation Attacks
Anthropic revealed that three Chinese AI companies — DeepSeek, Moonshot AI, and MiniMax — orchestrated industrial-scale distillation campaigns using...
- News
CrowdStrike 2026 Threat Report: eCrime Breakout Time Falls
CrowdStrike's 2026 Global Threat Report reveals that AI-enabled adversary operations surged 89% year-over-year, the average eCrime breakout time dropped...
- News
Alibaba Launches Qwen 3.5 — Claims to Outperform GPT-5.2
Alibaba Cloud unveils Qwen 3.5, a 397-billion-parameter Mixture-of-Experts AI model with native multimodal capabilities, claiming 60% lower costs and...
- News
Anthropic Hits $380 Billion Valuation After Closing $30
Anthropic more than doubles its valuation to $380 billion following a $30 billion funding round, intensifying the AI race with OpenAI amid the launch of...
- News
India Hosts Global AI Impact Summit — 20 World Leaders and Tech CEOs
India's AI Impact Summit kicks off February 16 in New Delhi, drawing 20 national leaders, 45 ministerial delegations, and the CEOs of Anthropic, OpenAI,...
- News
Claude AI Artifacts Abused to Distribute macOS Infostealer
Threat actors are abusing publicly shared Claude AI artifacts and Google Ads to deliver the MacSync infostealer to macOS users through ClickFix social...
- News
AI Chat App Exposes 300 Million Private Messages from 25
A misconfigured Google Firebase backend in the Chat & Ask AI app exposed 300 million private chatbot conversations from 25 million users, including...
- News
All Four Major Nation-State Adversaries Now Weaponizing
Google reports that APT groups from China, Russia, Iran, and North Korea are all actively using Gemini AI for cyber operations including target...
- News
North Korea Deploys AI-Generated Video and ClickFix
North Korean threat actors are running sophisticated campaigns using AI-generated deepfake videos and the ClickFix social engineering technique to target...
- HOWTO
SentinelOne Control vs Complete Feature Comparison
This document provides a comprehensive comparison between SentinelOne Singularity Control and Singularity Complete SKUs to help MSP teams understand the...
- HOWTO
SentinelOne Policy Configuration Best Practices
This guide provides comprehensive best practices for configuring SentinelOne policies in MSP environments managing multiple client sites with Singularity...
- HOWTO
SentinelOne Purple AI Usage Guide
Security Operations Centers (SOCs) face overwhelming alert volumes, complex threat investigations, and resource constraints. Analysts spend hours writing...
- News
VoidLink: AI-Generated Cloud-Native Malware Framework
Researchers uncover VoidLink, an 88,000-line Zig-based malware framework built with AI assistance that targets AWS, Azure, GCP, and Kubernetes environments.
- News
FBI Warns of AI-Generated Deepfake Phishing Targeting
The FBI and CISA issue joint advisory on sophisticated AI-generated deepfake voice and video attacks targeting C-suite executives in financial...
- Security
International AI Safety Report: AI Now Provides 'Meaningful
The 2026 International AI Safety Report confirms AI systems can assist attackers across multiple stages of the cyberattack chain, with vulnerability...
- News
Amazon Alexa+ Goes GA After Tens of Millions Join Beta
Amazon's AI-powered Alexa+ assistant reaches general availability with enhanced conversational AI, smart home integration, and multi-modal capabilities.
- News
SpaceX-xAI Partnership Values Combined Entities Near $1.25
SpaceX's strategic tie-up with xAI and planned mid-2026 IPO could reshape the tech sector, with SpaceX valued near $1 trillion and xAI at $250 billion.
- News
AI-Powered Cyberattacks Expected to Cause Major Enterprise
Security experts predict autonomous AI systems will be responsible for at least one major enterprise breach within months, as threat actors weaponize...
- Project
Building a RAG System Without ML Embeddings
Create a lightweight, offline-capable documentation search system using keyword-based scoring instead of ML embeddings. Perfect for air-gapped...
- News
Microsoft Announces Major Security Features for Copilot
New AI-powered threat detection and automated incident response capabilities coming to Microsoft 365 E5 subscribers this quarter.
- News
AI-Powered Phishing Achieves 54% Click-Through Rate
Microsoft reveals adversaries using AI for automated vulnerability discovery, phishing campaigns, and malware generation. AI-crafted phishing emails...
- News
Cybersecurity Predictions 2026: The Hype We Can Ignore and the Real Risks
Industry experts separate signal from noise in 2026's threat landscape. AI-powered attacks, supply chain risks, and the evolution of ransomware top the...