An incident and vulnerability tracker for AI/ML systems. Model leaks, training-data exposures, jailbreak disclosures, ML library CVEs, vendor breaches, and confirmed prompt-injection-in-the-wild — each entry linked to a primary source, dated, and tagged for filtering.https://ai-alert.org/enhttps://ai-alert.org/posts/weekly-who-runs-the-ransomware-group-the-gentlemen/https://ai-alert.org/posts/weekly-who-runs-the-ransomware-group-the-gentlemen/Security researchers have identified the suspected administrator of The Gentlemen, the world's second-most-active RaaS by victim count, as a RussianFri, 12 Jun 2026 00:00:00 GMTransomwareraasthreat-intelligenceai-assisted-attackattributioncveAI Alert Deskhttps://ai-alert.org/posts/llm-security-risks-2/https://ai-alert.org/posts/llm-security-risks-2/A practitioner breakdown of LLM security risks covering the OWASP Top 10 for LLM Applications 2025, MITRE ATLAS AI attack tactics, and concrete mitigationWed, 03 Jun 2026 00:00:00 GMTllm-securityprompt-injectionowaspsupply-chainai-vulnerabilitiesAI Alert Deskhttps://ai-alert.org/posts/ai-security-incident-response-plan/https://ai-alert.org/posts/ai-security-incident-response-plan/A practical incident response plan for AI systems — what a prompt-injection or model-extraction incident looks like, how the NIST SP 800-61r3 / CSF 2.Sat, 23 May 2026 00:00:00 GMTincident-responsenistai-securityplaybookdetectioncontainmentforensicsAI Alert Deskhttps://ai-alert.org/posts/nist-ai-rmf-practitioner-field-guide/https://ai-alert.org/posts/nist-ai-rmf-practitioner-field-guide/What the NIST AI RMF actually asks you to do — the GOVERN, MAP, MEASURE, MANAGE functions, the Generative AI Profile's 12 risk categories, and how aSat, 23 May 2026 00:00:00 GMTnistai-rmfgovernancecompliancerisk-managementgenerative-aiframeworkAI Alert Deskhttps://ai-alert.org/posts/cisa-admin-leaked-aws-govcloud-keys-on-github/https://ai-alert.org/posts/cisa-admin-leaked-aws-govcloud-keys-on-github/A Nightwing contractor's public GitHub repository exposed administrative credentials to three AWS GovCloud accounts and dozens of internal CISA systemsTue, 19 May 2026 00:00:00 GMTcredential-leakawsgovcloudgithubcisaAI Alert Deskhttps://ai-alert.org/posts/deepfake-cybersecurity-3/https://ai-alert.org/posts/deepfake-cybersecurity-3/From the FBI's May 2025 warning on AI voice attacks targeting US officials to NIST's synthetic content framework, here is what detection technologyThu, 14 May 2026 00:00:00 GMTdeepfakessynthetic-mediadetectionvishingnistai-securityAI Alert Deskhttps://ai-alert.org/posts/generative-ai-risks-2/https://ai-alert.org/posts/generative-ai-risks-2/A practitioner-focused breakdown of generative AI risks mapped against NIST AI 600-1 and the OWASP Top 10 for LLMs — prompt injection, data poisoningThu, 14 May 2026 00:00:00 GMTgenerative-aillm-securityprompt-injectiondata-poisoningai-risk-managementAI Alert Deskhttps://ai-alert.org/posts/machine-learning-security-3/https://ai-alert.org/posts/machine-learning-security-3/A technical overview of machine learning security threats in 2026: NIST's adversarial ML taxonomy, MITRE ATLAS attack classes, the CVE-2025-62164 vLLMWed, 13 May 2026 00:00:00 GMTmachine-learningadversarial-mlcvedata-poisoningsupply-chainmodel-securityAI Alert Deskhttps://ai-alert.org/posts/openai-security/https://ai-alert.org/posts/openai-security/A practitioner's overview of OpenAI security in 2026: their bug bounty program, CNA status, the November 2025 Mixpanel breach, and what security teamsTue, 12 May 2026 00:00:00 GMTopenaibug-bountycvebreachvulnerability-disclosureai-securityAI Alert Deskhttps://ai-alert.org/posts/chatgpt-security-2/https://ai-alert.org/posts/chatgpt-security-2/A technical review of ChatGPT security vulnerabilities disclosed in 2025-2026: DNS-based data exfiltration, ZombieAgent prompt injection bypass, CodexMon, 11 May 2026 00:00:00 GMTchatgptopenaiprompt-injectiondata-exfiltrationllm-securityvulnerability-disclosureAI Alert Deskhttps://ai-alert.org/posts/chatgpt-security/https://ai-alert.org/posts/chatgpt-security/A practitioner's guide to ChatGPT security in 2026: how OpenAI protects enterprise data, where prompt injection and account-takeover risks live, and theMon, 11 May 2026 00:00:00 GMTchatgptllm-securityopenaiprompt-injectionenterprise-aidata-privacyAI Alert Deskhttps://ai-alert.org/posts/cisa-ai-security-guidance-2026/https://ai-alert.org/posts/cisa-ai-security-guidance-2026/A breakdown of CISA's published AI security guidance — what it covers, what it requires, and how organizations should operationalize it.Mon, 11 May 2026 00:00:00 GMTcisaai-security-guidancesecure-by-designai-sbomcomplianceadvisoryfederalthreat-alertAI Alert Deskhttps://ai-alert.org/posts/deepfake-cybersecurity-2/https://ai-alert.org/posts/deepfake-cybersecurity-2/A working catalog of confirmed deepfake cybersecurity incidents from 2024 and 2025 — from Arup's $25M loss to Ferrari's averted scam — and the controlsMon, 11 May 2026 00:00:00 GMTdeepfakesincident-catalogsocial-engineeringbecdprk-it-workersai-securityAI Alert Deskhttps://ai-alert.org/posts/deepfake-cybersecurity/https://ai-alert.org/posts/deepfake-cybersecurity/Deepfake cybersecurity has moved from theoretical risk to documented billion-dollar loss category. Here's what the attack surface looks like, whyMon, 11 May 2026 00:00:00 GMTdeepfakessynthetic-mediasocial-engineeringidentity-fraudbecthreat-briefai-securityAI Alert Deskhttps://ai-alert.org/posts/generative-ai-risks/https://ai-alert.org/posts/generative-ai-risks/Generative AI risks span prompt injection, data poisoning, supply chain vulnerabilities, hallucination, and governance failures.Mon, 11 May 2026 00:00:00 GMTgenerative-aiai-risksllm-securityprompt-injectiondata-poisoningai-governanceAI Alert Deskhttps://ai-alert.org/posts/how-to-track-ai-security-alerts/https://ai-alert.org/posts/how-to-track-ai-security-alerts/A practical guide to the official and community sources for AI security alerts — what each publishes, how frequently, and how to integrate them into aMon, 11 May 2026 00:00:00 GMTai-security-alertcisanistadvisorymonitoringthreat-intelligencepatch-managementAI Alert Deskhttps://ai-alert.org/posts/llm-security-alerts-monitoring-response/https://ai-alert.org/posts/llm-security-alerts-monitoring-response/A practical guide to setting up LLM security alerting — what to monitor, what alert patterns indicate compromise or attack, how to triage LLM securityMon, 11 May 2026 00:00:00 GMTllm-security-alertai-security-alertmonitoringdetectionincident-responseprompt-injectionai-security-updatesiemAI Alert Deskhttps://ai-alert.org/posts/llm-security-risks/https://ai-alert.org/posts/llm-security-risks/A comprehensive breakdown of LLM security risks — prompt injection, supply chain poisoning, excessive agency, and model extraction — with mitigationMon, 11 May 2026 00:00:00 GMTllm-securityprompt-injectionsupply-chainmodel-poisoningagent-securityowaspAI Alert Deskhttps://ai-alert.org/posts/machine-learning-security-2/https://ai-alert.org/posts/machine-learning-security-2/Machine learning security vulnerabilities enter at every stage — data ingestion, model training, artifact storage, and inference.Mon, 11 May 2026 00:00:00 GMTmachine-learning-securityml-pipelinedata-poisoningsupply-chainmlsecopsAI Alert Deskhttps://ai-alert.org/posts/machine-learning-security/https://ai-alert.org/posts/machine-learning-security/A practitioner's reference for machine learning security: the canonical attack categories, the frameworks that catalog them (NIST AI 100-2, OWASP ML TopMon, 11 May 2026 00:00:00 GMTmachine-learning-securityadversarial-mlmlsecopsnistowaspmitre-atlasAI Alert Deskhttps://ai-alert.org/posts/ai-generated-phishing-spearphishing-cost-collapse/https://ai-alert.org/posts/ai-generated-phishing-spearphishing-cost-collapse/Crafting a convincing, personalized phishing email once required hours of research per target. Large language models have reduced that cost to seconds.Sun, 10 May 2026 00:00:00 GMTphishingspearphishingsocial-engineeringllm-abusethreat-intelligencebecenterprise-securityAI Alert Deskhttps://ai-alert.org/posts/ai-agent-security-incidents-2025/https://ai-alert.org/posts/ai-agent-security-incidents-2025/A documented review of security incidents involving autonomous AI agents in 2024-2025, covering tool misuse, privilege escalation via injection, and theSun, 10 May 2026 00:00:00 GMTagent-securityagentic-aillm-agentsprompt-injectiontool-useincidentsautonomous-aiAI Alert Deskhttps://ai-alert.org/posts/ai-red-teaming-llm-jailbreak-defenses-2026/https://ai-alert.org/posts/ai-red-teaming-llm-jailbreak-defenses-2026/Enterprise LLM deployments are being red-teamed at scale for the first time. Security practitioners find consistent failure patterns — misconfiguredSun, 10 May 2026 00:00:00 GMTred-teamjailbreakllm-securitysystem-promptagenticenterprise-aiassessment2026AI Alert Deskhttps://ai-alert.org/posts/cisa-kev-ai-ml-vulnerabilities-analysis/https://ai-alert.org/posts/cisa-kev-ai-ml-vulnerabilities-analysis/The CISA KEV catalog tracks vulnerabilities with confirmed active exploitation. Examining KEV entries for AI/ML-adjacent components reveals which parts ofSun, 10 May 2026 00:00:00 GMTcisakevknown-exploitedvulnerabilityml-infrastructurecvepatch-managementfederalAI Alert Deskhttps://ai-alert.org/posts/compromised-huggingface-models-pickle-exploits/https://ai-alert.org/posts/compromised-huggingface-models-pickle-exploits/Malicious actors have uploaded model files to Hugging Face containing pickle payloads that execute code on download.Sun, 10 May 2026 00:00:00 GMThugging-facepicklemodel-hubsupply-chaincode-executionsafetensorsmalicious-modelsmlsecAI Alert Deskhttps://ai-alert.org/posts/cve-roundup-ai-ml-infrastructure-q1-2026/https://ai-alert.org/posts/cve-roundup-ai-ml-infrastructure-q1-2026/A quarterly review of critical CVEs disclosed in Q1 2026 affecting model serving infrastructure: vLLM, NVIDIA Triton Inference Server, Gradio, LangChainSun, 10 May 2026 00:00:00 GMTcvecve-roundupvllmtritongradiolangchainlangflowmodel-servingq1-2026infrastructureAI Alert Deskhttps://ai-alert.org/posts/hugging-face-security-incidents/https://ai-alert.org/posts/hugging-face-security-incidents/A review of documented security incidents on the Hugging Face platform, including malicious model uploads, the 2024 Spaces infrastructure breach, and theSun, 10 May 2026 00:00:00 GMThugging-facesupply-chainmodel-hubpicklesafetensorsinfrastructure-securitymlopsAI Alert Deskhttps://ai-alert.org/posts/jailbreak-techniques-2025-roundup/https://ai-alert.org/posts/jailbreak-techniques-2025-roundup/A roundup of significant jailbreak techniques disclosed or widely documented in 2025, including many-shot jailbreaking, crescendo attacks, cipher-basedSun, 10 May 2026 00:00:00 GMTjailbreakingalignment-bypassred-teamingllm-securitydisclosuresafetyadversarial-promptingAI Alert Deskhttps://ai-alert.org/posts/model-file-format-vulnerabilities-pickle-onnx-safetensors/https://ai-alert.org/posts/model-file-format-vulnerabilities-pickle-onnx-safetensors/Unsafe deserialization in PyTorch's pickle-based format has enabled malicious model distribution for years. This post explains how pickle exploitationSun, 10 May 2026 00:00:00 GMTpickleonnxsafetensorsmodel-formatdeserializationsupply-chainhuggingfacepytorchAI Alert Deskhttps://ai-alert.org/posts/owasp-llm-top-10-2025-changes/https://ai-alert.org/posts/owasp-llm-top-10-2025-changes/The OWASP Top 10 for Large Language Model Applications was updated for 2025. Here is a breakdown of what moved, what was added, and why the changesSun, 10 May 2026 00:00:00 GMTowaspllm-securitytop-10prompt-injectionragagent-securityrisk-frameworkAI Alert Deskhttps://ai-alert.org/posts/prompt-injection-email-ai-agents/https://ai-alert.org/posts/prompt-injection-email-ai-agents/Email is the highest-volume source of untrusted content in enterprise environments — and it's now being fed directly into AI agents.Sun, 10 May 2026 00:00:00 GMTprompt-injectionemailcopilotoutlookindirect-injectionmicrosoft-365enterprise-aiincidentAI Alert Deskhttps://ai-alert.org/posts/rag-knowledge-base-poisoning-2026-threat-brief/https://ai-alert.org/posts/rag-knowledge-base-poisoning-2026-threat-brief/Attackers are actively poisoning retrieval-augmented generation knowledge bases in enterprise deployments. This briefing documents the current threatSun, 10 May 2026 00:00:00 GMTragdata-poisoningknowledge-basevector-databaseindirect-injectionenterprise-aithreat-brief2026AI Alert Deskhttps://ai-alert.org/posts/rag-poisoning-retrieval-attacks/https://ai-alert.org/posts/rag-poisoning-retrieval-attacks/RAG systems inherit all the vulnerabilities of LLMs and add a new one: the retrieval corpus. Injecting malicious content into retrieved sources can hijackSun, 10 May 2026 00:00:00 GMTragretrievalprompt-injectionindirect-injectionvector-databaseknowledge-basellm-securityAI Alert Deskhttps://ai-alert.org/posts/shadow-ai-enterprise-ungoverned-api-keys/https://ai-alert.org/posts/shadow-ai-enterprise-ungoverned-api-keys/Employees using personal Claude, OpenAI, and Gemini API keys for work tasks bypass corporate DLP controls and send sensitive business data to externalSun, 10 May 2026 00:00:00 GMTshadow-aiapi-keysdata-exfiltrationenterprise-securitydlpgovernanceopenaiclaudeAI Alert Deskhttps://ai-alert.org/posts/ai-security/https://ai-alert.org/posts/ai-security/A practitioner guide to the four core attack categories against AI/ML systems — from adversarial inputs to supply chain compromise — with mitigationSat, 09 May 2026 00:00:00 GMTai-securityadversarial-mlprompt-injectionsupply-chainnistllm-securityAI Alert Deskhttps://ai-alert.org/posts/model-extraction-attacks-explained/https://ai-alert.org/posts/model-extraction-attacks-explained/Model extraction attacks reconstruct proprietary AI models by querying their public APIs. Here's how they work, what has been demonstrated against realSat, 09 May 2026 00:00:00 GMTmodel-extractionmodel-theftapi-securityadversarial-mlintellectual-propertyprivacyllm-securityAI Alert Deskhttps://ai-alert.org/posts/weekly-ai-security-digest-may-week2/https://ai-alert.org/posts/weekly-ai-security-digest-may-week2/Top five AI security developments from May 5-9, 2026: CISA guidance on AI in critical infrastructure, new prompt injection research, LLM supply chainSat, 09 May 2026 00:00:00 GMTdigestweeklycisaprompt-injectionsupply-chainbreacheu-ai-actmay-2026AI Alert Deskhttps://ai-alert.org/posts/ai-red-teaming-guide/https://ai-alert.org/posts/ai-red-teaming-guide/Red-teaming LLMs requires different skills and methodology than traditional network or application penetration testing.Fri, 08 May 2026 00:00:00 GMTred-teamingllm-securitypentestingadversarial-testingmethodologyowaspai-securityAI Alert Deskhttps://ai-alert.org/posts/ai-security-audit-checklist/https://ai-alert.org/posts/ai-security-audit-checklist/A practical audit checklist for AI systems covering model inputs, training pipeline, outputs, access control, logging, and red-team requirements.Fri, 08 May 2026 00:00:00 GMTchecklistauditprompt-injectiondata-poisoningaccess-controlloggingred-teamprimerAI Alert Deskhttps://ai-alert.org/posts/aisecbench-2026-evaluation-methodology/https://ai-alert.org/posts/aisecbench-2026-evaluation-methodology/Choosing an AI security tool without a structured evaluation methodology is expensive guesswork. This guide covers the metrics that matter, the pitfallsFri, 08 May 2026 00:00:00 GMTbenchmarkingevaluationmethodologyai-security-toolsred-teamingmlsecopsmetricsAI Alert Deskhttps://ai-alert.org/posts/ai-incident-samsung-data-leak/https://ai-alert.org/posts/ai-incident-samsung-data-leak/In 2023, Samsung employees leaked proprietary source code and meeting notes through ChatGPT. This incident defined a category of enterprise AI risk thatThu, 07 May 2026 00:00:00 GMTdata-leaksamsungchatgptenterprise-aidlpincident-analysisai-policyAI Alert Deskhttps://ai-alert.org/posts/prompt-injection-wild-2025/https://ai-alert.org/posts/prompt-injection-wild-2025/A catalog of confirmed prompt injection incidents in real deployments: Bing Chat, Slack AI, email assistants, and customer service bots.Thu, 07 May 2026 00:00:00 GMTprompt-injectionbingslackincidentsanalysisindirect-injectionwildAI Alert Deskhttps://ai-alert.org/posts/tool-review-garak-llm-scanner/https://ai-alert.org/posts/tool-review-garak-llm-scanner/Garak is an open-source LLM vulnerability scanner from NVIDIA that probes language models for dozens of failure modes.Thu, 07 May 2026 00:00:00 GMTgaraktool-reviewllm-scannervulnerability-scannerred-teamingopen-sourcenvidiaAI Alert Deskhttps://ai-alert.org/posts/cve-2026-7845-a-flaw-has-been-found-in-chatchat-space-langch/https://ai-alert.org/posts/cve-2026-7845-a-flaw-has-been-found-in-chatchat-space-langch/A weak-hash flaw in Langchain-Chatchat up to 0.3.1.3 lets an adjacent attacker overwrite pasted images by colliding MD5 hashes computed from PIL.Image.Wed, 06 May 2026 00:00:00 GMTcvelangchain-chatchatweak-hashvision-llmmd5image-collisionAI Alert Deskhttps://ai-alert.org/posts/cve-roundup-ai-ml-may-2026/https://ai-alert.org/posts/cve-roundup-ai-ml-may-2026/A summary of AI and ML-adjacent CVEs disclosed in early–mid 2026 across model serving frameworks, LLM API gateways, agent SDKs, and ML training libraries.Wed, 06 May 2026 00:00:00 GMTcvecve-roundupvllmlitellmollamasupply-chainmay-2026AI Alert Deskhttps://ai-alert.org/posts/gpt4-model-theft-incident-analysis/https://ai-alert.org/posts/gpt4-model-theft-incident-analysis/Model extraction attacks let adversaries reconstruct proprietary LLMs by querying their APIs at scale. We examine the mechanics, cost-of-extractionWed, 06 May 2026 00:00:00 GMTmodel-extractionmodel-theftinference-attacksopenaianthropicapi-securityanalysisAI Alert Deskhttps://ai-alert.org/posts/llm-supply-chain-poisoning/https://ai-alert.org/posts/llm-supply-chain-poisoning/Training data poisoning and model supply chain attacks are among the hardest AI threats to detect. This post explains how they work, what public researchWed, 06 May 2026 00:00:00 GMTsupply-chainpoisoningbackdoortraining-datallm-securityadversarial-mlhuggingfaceAI Alert Deskhttps://ai-alert.org/posts/ai-tool-security-review-llm-guard/https://ai-alert.org/posts/ai-tool-security-review-llm-guard/LLM Guard is an open-source input/output filtering library for LLM applications. We review what it detects, how it deploys, its real limitations, and whenTue, 05 May 2026 00:00:00 GMTtoolsllm-guardinput-filteringoutput-filteringprompt-injectionpiidefenseAI Alert Deskhttps://ai-alert.org/posts/gpt4-system-prompt-leak-incident/https://ai-alert.org/posts/gpt4-system-prompt-leak-incident/Prompt injection attacks that expose system prompts are one of the most common real-world LLM exploits. This post covers the mechanics of system promptTue, 05 May 2026 00:00:00 GMTprompt-injectionsystem-promptllm-securitygpt-4jailbreakingowaspconfidentialityAI Alert Deskhttps://ai-alert.org/posts/jailbreaking-vs-prompt-injection/https://ai-alert.org/posts/jailbreaking-vs-prompt-injection/Security practitioners conflate jailbreaking and prompt injection constantly. They are distinct attack classes with different threat actors, differentTue, 05 May 2026 00:00:00 GMTjailbreakingprompt-injectionllm-securityowaspalignmentprimerAI Alert Deskhttps://ai-alert.org/posts/metinfo-cms-cve-2026-29014-exploited-for-remote-code-executi/https://ai-alert.org/posts/metinfo-cms-cve-2026-29014-exploited-for-remote-code-executi/A critical unauthenticated PHP code injection flaw in MetInfo CMS 7.9–8.1 (CVSS 9.8) is under active exploitation. Patch to the April 7 release immediately.Tue, 05 May 2026 00:00:00 GMTcveremote-code-executionphpcmsactive-exploitationAI Alert Deskhttps://ai-alert.org/posts/weekly-germany-doxes-unkn-head-of-ru-ransomware-gangs-revil-gandcra/https://ai-alert.org/posts/weekly-germany-doxes-unkn-head-of-ru-ransomware-gangs-revil-gandcra/Germany's BKA has put a name and a face to UNKN, the operator behind GandCrab and REvil. Russia will not extradite, but the wanted notice is doing otherTue, 05 May 2026 00:00:00 GMTransomwarerevilgandcrabattributionthreat-intellaw-enforcementAI Alert Deskhttps://ai-alert.org/posts/what-is-adversarial-ml-primer/https://ai-alert.org/posts/what-is-adversarial-ml-primer/A practitioner-focused introduction to adversarial machine learning: evasion, poisoning, and inference attacks, why they matter in production, key papersMon, 04 May 2026 00:00:00 GMTadversarial-mlevasionpoisoningmembership-inferenceprimernistAI Alert Deskhttps://ai-alert.org/posts/cisa-adds-actively-exploited-linux-root-access-bug-cve-2026/https://ai-alert.org/posts/cisa-adds-actively-exploited-linux-root-access-bug-cve-2026/A local privilege escalation flaw in the Linux kernel's AEAD crypto interface has been added to CISA's KEV catalog after active exploitation.Sun, 03 May 2026 00:00:00 GMTcvelinux-kernellocal-privilege-escalationcisa-kevcryptographyAI Alert Deskhttps://ai-alert.org/posts/cve-2026-7669-a-vulnerability-was-detected-in-sgl-project-sg/https://ai-alert.org/posts/cve-2026-7669-a-vulnerability-was-detected-in-sgl-project-sg/A medium-severity deserialization bug in SGLang's get_tokenizer routine affects all releases up to 0.5.9. The vendor has not responded to the disclosureSun, 03 May 2026 00:00:00 GMTcvesglanghuggingfacedeserializationllm-servingsupply-chainAI Alert Desk