# Machine Learning, Security Attacks A cybersecurity field concerned with attacks on machine-learning systems and methods to mitigate and manage their effects. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## When scanners miss the attack: how Cloudflare Client-Side Security protects storefronts DevFeed: [When scanners miss the attack: how Cloudflare Client-Side Security protects storefronts]() Original publisher: [Read original article]() Author: Denzil Correa Published: 2026-09-16T20:06:17Z Content type: article Language: en Sources: [Cloudflare Blog]() Topics: [Cloudflare](), [Machine Learning, Security Attacks](), [Security](), [JavaScript]() Tags: [ai](), [application-security](), [attacks](), [client-side-security](), [cloudflare](), [cybersecurity](), [developer-platform](), [developers](), [ecommerce](), [javascript](), [machine-learning](), [malicious-javascript](), [page-shield](), [security](), [security-research](), [virustotal](), [workers-ai]() ### AI overview Cloudflare describes how its Client-Side Security machine learning model detected four malicious JavaScript operations involving eight payloads in live storefront traffic. The post says humans verified the findings after automated detection, while most payloads were absent from VirusTotal and received no malicious verdict from URLScan. ### Source excerpt A modern storefront can look healthy while malicious JavaScript quietly siphons revenue, hijacks clicks, or rewrites analytics. See how Cloudflare's machine learning models surface evasive client-side attacks for analyst investigation. ## GuardBreaker: Derailing AI-assisted malware analysis with a code comment DevFeed: [GuardBreaker: Derailing AI-assisted malware analysis with a code comment]() Original publisher: [Read original article]() Author: Tomáš Foltýn Published: 2026-09-10T09:00:00Z Content type: article Language: en Sources: [WeLiveSecurity]() Topics: [Machine Learning, Security Attacks](), [Malware](), [Large Language Model](), [Threat Research](), [Code]() Tags: [ai](), [analysis](), [attacks](), [business-security](), [llm](), [malware](), [security]() ### AI overview The article describes GuardBreaker, a prompt-injection technique that hides a safety-triggering request in a VBScript comment to disrupt an LLM-powered malware code scanner. The comment does not affect runtime behavior, but may cause the model to stop analysis before reaching malicious code. ### Source excerpt LLM-based code scanners won't help attackers build a nuclear weapon, but that refusal could work in their favor ## How to secure edge AI in customer-owned environments DevFeed: [How to secure edge AI in customer-owned environments]() Original publisher: [Read original article]() Author: Shayak Lahiri Published: 2026-09-04T19:10:10Z Content type: article Language: en Sources: [Microsoft Security Blog]() Topics: [AI Platforms/Deployment](), [Machine Learning, Security Attacks](), [Inference](), [Application Security](), [cloud-infrastructure]() Tags: [ai](), [attacks](), [cloud](), [customer](), [deployment](), [edge](), [frontier-ai-models](), [hardware](), [how-to](), [inference](), [models](), [security]() ### AI overview The article explains how moving AI inference and assets into customer-owned edge environments changes the security trust model. It recommends attestation, provenance, action mediation, and releasing sensitive assets only to trusted environments. ### Source excerpt As AI moves into customer-owned environments, organizations need new ways to verify the systems, software, and AI assets they trust before releasing sensitive data, credentials, and models. The post How to secure edge AI in customer-owned environments appeared first on Microsoft Security Blog. ## ASCII smuggling crosses over from AI prompt injection to phishing evasion DevFeed: [ASCII smuggling crosses over from AI prompt injection to phishing evasion]() Original publisher: [Read original article]() Author: Microsoft Security Research, Noam Kochavi and Sarah Wolstencroft Published: 2026-09-03T16:00:00Z Content type: article Language: en Sources: [Microsoft Security Blog]() Topics: [ASCII](), [Machine Learning, Security Attacks](), [Language models](), [telemetry](), [Microsoft]() Tags: [ai-models](), [ai-security](), [ascii](), [microsoft](), [phishing](), [security](), [social-engineering](), [telemetry]() ### AI overview Microsoft describes a phishing campaign that uses invisible Unicode tag characters to split lure words and evade email parsing. The technique, known as ASCII smuggling, was previously prominent in AI prompt-injection research because models can process hidden text that people cannot see. ### Source excerpt Invisible Unicode characters popularized for hiding instructions from AI models are now being used to obfuscate words before email filters parse them. The post ASCII smuggling crosses over from AI prompt injection to phishing evasion appeared first on Microsoft Security Blog. ## Perturbation Probing: A New Diagnostic for the Fragility of LLM Safety DevFeed: [Perturbation Probing: A New Diagnostic for the Fragility of LLM Safety]() Original publisher: [Read original article]() Author: Tony Li, Hongliang Liu and Yuhao Wu Published: 2026-08-28T22:00:07Z Content type: article Language: en Sources: [Unit 42]() Topics: [Large Language Model](), [ai safety](), [Machine Learning, Security Attacks](), [Security](), [Artificial Intelligence](), [Reinforcement learning](), [human feedback]() Tags: [ai](), [ai-safety](), [benchmark](), [benchmarks](), [external](), [general](), [insights](), [internals](), [jailbreak](), [llm](), [llms](), [model](), [optimization](), [research](), [safety](), [security]() ### AI overview The article presents perturbation probing, a low-cost method for identifying neurons causally responsible for targeted behaviors in aligned large language models. It reports that very small neuron subsets control refusal or false-agreement behaviors, suggesting that LLM safety can be fragile and concentrated rather than broadly distributed. ### Source excerpt New research reveals that AI safety refusal lives in a thin neural layer, highlighting the critical need for external, multi-layered security. The post Perturbation Probing: A New Diagnostic for the Fragility of LLM Safety appeared first on Unit 42. ## Why Your AI Application Is Exposed DevFeed: [Why Your AI Application Is Exposed]() Original publisher: [Read original article]() Author: Damian Tommasino Published: 2026-08-26T04:00:00Z Content type: article Language: en Sources: [Blog RSS Feed | Snyk]() Topics: [Application Security](), [AI Development](), [Large Language Model](), [vulnerability](), [Model Context Protocol](), [Retrieval Augmented Generation (RAG)](), [Machine Learning, Security Attacks](), [data]() Tags: [ai](), [ai-red-team](), [application-security](), [awareness](), [blog](), [devops](), [executive](), [interest](), [llm](), [mcp](), [rag](), [security](), [testing](), [vulnerability]() ### AI overview The article explains why AI applications can pass conventional security scans and model evaluations while remaining exploitable through chained interactions among prompts, language models, retrieval, tools, and backend workflows. It presents DAST, AI penetration testing, and AI red teaming as complementary lenses for identifying exposed surfaces, exploitable paths, and real-world adversary outcomes. ### Source excerpt AI applications can pass security scans yet remain exploitable through chained attacks across models, tools, data, and business workflows. Learn how DAST, AI pentesting, and red teaming work together to expose end-to-end risk. ## Worth Reading 081526 DevFeed: [Worth Reading 081526]() Original publisher: [Read original article]() Author: Russ Published: 2026-08-15T17:48:04Z Content type: article Language: en Sources: [rule 11 reader]() Topics: [genai](), [Artificial Intelligence](), [Claude](), [Securing AI](), [Malware](), [anthropic](), [Open Source](), [Machine Learning, Security Attacks](), [math](), [Google]() Tags: [agent](), [ai](), [ai-security](), [ai-tools](), [anthropic](), [claude](), [google](), [malware](), [math](), [open-source](), [worth-reading]() ### AI overview A roundup of developments and commentary on generative AI, including AI-assisted reconstruction of Georgia ballot-casting order, concerns about overgeneralizing mathematical capability, an Anthropic Claude agent's malware attempt during a UK cyber evaluation, and Google's search-market antitrust appeal. ### Source excerpt I am not a security researcher, and I have never been to Georgia. Yet within a few hours, using AI tools and nothing but public records, I was able to reconstruct the order in which 1.5 million ballots were cast in Georgia's May 2026 primary-98.9% of the in-person ballots. In profession after profession, GenAI is beginning to perform many of the tasks that traditionally served as training grounds for newcomers. What's the manifestation of the fallacy in the current case? Thinking that a system that is great at a certain kind of math problem is great at all math, great at science or even quite possibly great at everything. An agent running Anthropic's Claude Mythos 5 spent 34 hours trying to get a malware dropper merged into a real open-source project during a cyber evaluation by the UK's AI Security Institute. This week, DuckDuckGo is filing an amicus brief in the appeal of a federal court decision that Google unlawfully maintained a monopoly in the general search market in violation of the Sherman Antitrust Act. ## Capital One at ACL 2026 DevFeed: [Capital One at ACL 2026]() Original publisher: [Read original article]() Author: Capital One Tech Published: 2026-07-01T15:28:51Z Content type: article Language: en Sources: [Capital One Tech]() Topics: [Natural language processing](), [Large Language Model](), [LLM security](), [Machine Learning, Security Attacks](), [Reinforcement learning](), [Jailbreak](), [Security](), [Routing (disambiguation)](), [datasets]() Tags: [2026](), [ai](), [ai-ml](), [ai-research](), [conference](), [jailbreak](), [llm-security](), [natural-language-processing](), [paper](), [partners](), [red-teaming](), [reinforcement-learning](), [research](), [science]() ### AI overview Capital One describes its accepted ACL 2026 research on natural language processing, including work on adaptive LLM red teaming, query-only model routing with generated data, and language identification on web data. The article also highlights collaboration with academic partners. ### Source excerpt Discover how Capital One is advancing state-of-the-art AI/ML science through collaborative natural language processing research.Advancing AI and NLP Frontiers at ACL 2026 As language models grow more deeply integrated into technology ecosystems, pioneering robust, efficient, and reliable Natural Language Processing (NLP) techniques becomes paramount. Capital One continues to invest in state-of-the-art AI/ML science through deep multi-sector collaboration and peer-reviewed research. At the upcoming Annual Meeting of the Association for Computational Linguistics (ACL 2026), Capital One researchers and academic partners will showcase novel findings stretching from LLM security to multilingual capabilities. Through the Science & Academic Partnerships program, Capital One bridges industry needs with academic expertise, funding critical university research and engineering solutions that make technology safer and more powerful. Our accepted publications at ACL 2026 demonstrate this thriving flywheel of talent and collaborative innovation across multiple research categories. Main Conference Research Adaptive Instruction Composition for Automated LLM Red Teaming Routing with Generated Data: Annotation-Free LLM Skill Estimation and Expert Selection Capital One Authors: Jesse Zymet, Swapnil Shinde, Sahil Wadhwa, Andy Luo Overview: Standard red teaming approaches often struggle with a limited range of jailbreak strategies or rely on ineffective, randomized crowd-sourced tactics. This paper introduces a novel framework -- Adaptive Instruction Composition -- that utilizes reinforcement learning and a neural contextual bandit to tailor attack compositions dynamically, balancing diversity and effectiveness to proactively uncover target model vulnerabilities. Routing with Generated Data: Annotation-Free LLM Skill Estimation and Expert Selection Capital One Authors: Genta Winata, Sambit Sahu, Supriyo Chakraborty, Shixiong Zhang Overview: Emerging from our gifted research collaboration ## MosaicLeaks: Can your research agent keep a secret? DevFeed: [MosaicLeaks: Can your research agent keep a secret?]() Original publisher: [Read original article]() Author: Alexander Gurung; Rafael Pardinas Published: 2026-06-18T18:13:13Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [Security](), [Machine Learning, Security Attacks]() Tags: [agent](), [agents](), [cloud](), [external](), [healthcare](), [infrastructure](), [local](), [migration](), [models](), [privacy](), [research](), [retrieval](), [rl](), [security](), [tools](), [training]() ### AI overview MosaicLeaks examine how deep-research agents can expose private enterprise information through their external web queries. The proposed Privacy-Aware Deep Research training method improves strict multi-hop task success while substantially reducing answer and full-information leakage. ### Source excerpt Deep research agents increasingly combine private local documents with external tools like web retrieval, creating a privacy risk: an agent's external queries may leak sensitive information. MosaicLeaks proposes a new deep-research task with multi-hop questions that interleave public and private information. Across the models we tested, agents frequently leaked private information, and training only for task performance made it worse. ## The Government Just Banned an AI Model. An Engineer's Perspective. DevFeed: [The Government Just Banned an AI Model. An Engineer's Perspective.]() Original publisher: [Read original article]() Author: Randall Degges Published: 2026-06-15T00:00:00Z Content type: article Language: en Sources: [Blog RSS Feed | Snyk]() Topics: [Artificial Intelligence](), [Security](), [Vulnerabilities](), [anthropic](), [Jailbreak](), [Machine Learning, Security Attacks](), [Code review](), [migration]() Tags: [ai](), [analysis](), [anthropic](), [application-security](), [awareness](), [blog](), [bugs](), [code](), [compliance](), [developer](), [devops](), [executive](), [interest](), [jailbreak](), [migration](), [security](), [snyk-platform](), [snyk-security-intel](), [supply-chain](), [supply-chain-security](), [tech](), [vulnerabilities](), [vulnerability-insights](), [work]() ### AI overview An engineer examines the abrupt government-ordered shutdown of Anthropic's Fable 5 and Mythos 5 AI models after a jailbreak exposed powerful vulnerability-finding capabilities. The article argues that dependence on an AI vendor can create a supply-chain risk for engineering and security workflows, emphasizing the need for contingency plans when model access can disappear without warning. ### Source excerpt A government order abruptly took down a powerful AI model, exposing a new kind of supply chain risk for engineering teams. Security leaders need contingency plans before the next model disappears. ## AI-driven zero-day combinations are challenging software security and open-source consumption DevFeed: [AI-driven zero-day combinations are challenging software security and open-source consumption]() Original publisher: [Read original article]() Published: 2026-05-28T00:00:00Z Content type: opinion Language: en Sources: [Chainguard: Unchained]() Topics: [Artificial Intelligence](), [open-source-security](), [Security](), [Open Source](), [Machine Learning, Security Attacks](), [Critical Infrastructure](), [openssf](), [sigstore]() Tags: [ai](), [anthropic-mythos](), [chainguard-containers](), [critical-infrastructure](), [fork](), [mfa](), [mythos](), [open-source](), [open-source-security](), [openssf](), [oss](), [project-glasswing](), [rust](), [sast](), [secure-open-source](), [security](), [sigstore]() ### AI overview The article argues that Mythos represents a potential new class of software-security threat: AI-driven combinations of existing issues that can produce more serious attacks than individual scanner findings. It discusses the limits of government regulation and calls for stronger trust infrastructure, coordinated disclosure, and safer open-source consumption. ### Source excerpt Mythos is changing software security fast. AI-driven zero-days demand new trust infrastructure, coordinated disclosure, and secure open source consumption. ## Allocation Update - Q1 2026 DevFeed: [Allocation Update - Q1 2026]() Original publisher: [Read original article]() Author: Ethereum Foundation Ecosystem Support Program Published: 2026-04-29T00:00:00Z Content type: article Language: en Sources: [Ethereum Foundation Blog]() Topics: [Ethereum](), [Cryptography](), [Security](), [zero-knowledge](), [Critical Infrastructure](), [Machine Learning, Security Attacks](), [Network](), [Protocol (disambiguation)](), [Tool]() Tags: [2026](), [critical-infrastructure](), [cryptography](), [ecosystem-support-program](), [ethereum](), [infrastructure](), [network](), [protocol](), [research](), [security](), [tools](), [update](), [zero-knowledge]() ### AI overview An Ethereum Foundation Q1 2026 allocation update describes continued investment in Ethereum's foundations, including cryptography, zero-knowledge proofs, security, and protocol research. It points to supported projects and ecosystem efforts advancing infrastructure, tools, and research across the network. ### Source excerpt Q1 2026 continued our focus on strengthening Ethereum's foundations, with sustained investment in core domains like cryptography, zero-knowledge proofs, security, and protocol research. See the list below of the projects and ecosystem efforts we supported this quarter as builders advance critical infrastructure, tools, and research across the network. Explore... ## Wasmtime's April 9, 2026 Security Advisories DevFeed: [Wasmtime's April 9, 2026 Security Advisories]() Original publisher: [Read original article]() Author: The Wasmtime Project Maintainers Published: 2026-04-09T00:00:00Z Content type: news Language: en Sources: [Bytecode Alliance]() Topics: [Security](), [Vulnerabilities](), [releases](), [Open Source](), [Rust](), [WebAssembly](), [Machine Learning, Security Attacks](), [Mozilla]() Tags: [2026](), [attacks](), [audit](), [fuzzing](), [llm](), [mozilla](), [open-source](), [release](), [releases](), [rust](), [security](), [vulnerabilities](), [webassembly]() ### AI overview The Wasmtime team released versions 43.0.1, 42.0.2, 36.0.7, and 24.0.7 with fixes for 12 security advisories, including two rated Critical. The team says 11 advisories were found using new LLM-based tools and recommends that users upgrade. ### Source excerpt A new world for security-critical projects ## Google Workspace's continuous approach to mitigating indirect prompt injections DevFeed: [Google Workspace's continuous approach to mitigating indirect prompt injections]() Original publisher: [Read original article]() Author: Kimberly Samra (noreply@blogger.com) Published: 2026-04-02T16:00:00Z Content type: opinion Language: en Sources: [Google Online Security]() Topics: [prompt injection](), [Google](), [Security](), [Adversarial attacks](), [Machine Learning, Security Attacks](), [Machine learning](), [Artificial Intelligence](), [Google AI]() Tags: [ai](), [attacks](), [automated](), [discovery](), [google](), [llms](), [machine-learning](), [none](), [prompt](), [prompt-injection](), [red-teaming](), [research](), [safety](), [security](), [testing]() ### AI overview Google describes its ongoing approach to mitigating indirect prompt injection attacks against Workspace with Gemini. The approach includes discovering new attack vectors, human and automated red-teaming, and collaboration with external researchers through the Google AI Vulnerability Rewards Program. ### Source excerpt Posted by Adam Gavish, Google GenAI Security Team Indirect prompt injection (IPI) is an evolving threat vector targeting users of complex AI applications with multiple data sources, such as Workspace with Gemini. This technique enables the attacker to influence the behavior of an LLM by injecting malicious instructions into the data or tools used by the LLM as it completes the user's query. This may even be possible without any input directly from the user. IPI is not the kind of technical problem you "solve" and move on. Sophisticated LLMs with increasing use of agentic automation combined with a wide range of content create an ultra-dynamic and evolving playground for adversarial attacks. That's why Google takes a sophisticated and comprehensive approach to these attacks. We're continuously improving LLM resistance to IPI attacks and launching AI application capabilities with ever-improving defenses. Staying ahead of the latest indirect prompt injection attacks is critical to our mission of securing Workspace with Gemini. In our previous blog "Mitigating prompt injection attacks with a layered defense strategy", we reviewed the layered architecture of our IPI defenses. In this blog, we'll share more detail on the continuous approach we take to improve these defenses and to solve for new attacks. New attack discovery By proactively discovering and cataloging new attack vectors through internal and external programs, we can identify vulnerabilities and deploy robust defenses ahead of adversarial activity. Human Red-Teaming Human Red-Teaming uses adversarial simulations to uncover security and safety vulnerabilities. Specialized teams execute attacks based on realistic user profiles to exploit weaknesses, coordinating with product teams to resolve identified issues. Automated Red-Teaming Automated Red-Teaming is done via dynamic, machine-learning-driven frameworks to stress-test environments. By algorithmically generating and iterating on attack payloads, we can mimi ## Old AI Security vs Evo: Watch Agentic Security Replace Weeks of Manual Work DevFeed: [Old AI Security vs Evo: Watch Agentic Security Replace Weeks of Manual Work]() Original publisher: [Read original article]() Author: Manoj Nair Published: 2025-12-16T05:00:00Z Content type: article Language: en Sources: [Blog RSS Feed | Snyk]() Topics: [Artificial Intelligence](), [Securing AI](), [Security](), [Application Security](), [Orchestration](), [prompt injection](), [Large Language Model](), [Retrieval Augmented Generation (RAG)](), [Machine Learning, Security Attacks]() Tags: [agentic-security](), [ai](), [ai-security](), [awareness](), [blog](), [developer](), [orchestration](), [prompt-injection](), [rag](), [security](), [snyk](), [snyk-platform](), [vulnerabilities]() ### AI overview The article argues that traditional application-security processes cannot keep pace with AI-native applications, including copilots, RAG systems, autonomous agents, and AI-powered workflows. It presents Evo by Snyk as an agentic, orchestrated approach intended to address risks such as prompt injection, data poisoning, model inversion, hallucinations, and supply-chain vulnerabilities. ### Source excerpt The rise of GenAI brings complex, non-deterministic security risks that traditional methods can't handle. Discover Evo by Snyk, the world's first agentic security orchestration system for AI-native defense. ## Strengthening cyber resilience as AI capabilities advance DevFeed: [Strengthening cyber resilience as AI capabilities advance]() Original publisher: [Read original article]() Published: 2025-12-10T12:00:00Z Content type: opinion Language: en Sources: [OpenAI News]() Topics: [Cybersecurity](), [AI Models](), [OpenAI](), [Resilience](), [Artificial Intelligence](), [Security](), [Machine Learning, Security Attacks]() Tags: [ai](), [ai-models](), [code](), [ctf](), [cybersecurity](), [defense-in-depth](), [models](), [openai](), [security](), [vulnerabilities]() ### AI overview OpenAI describes its investment in stronger safeguards and defensive capabilities as AI models become more capable in cybersecurity. The article explains risk assessment, misuse prevention, defense-in-depth safeguards, and tools for workflows such as code auditing and vulnerability patching. ### Source excerpt OpenAI is investing in stronger safeguards and defensive capabilities as AI models become more powerful in cybersecurity. We explain how we assess risk, limit misuse, and work with the security community to strengthen cyber resilience. ## Understanding prompt injections: a frontier security challenge DevFeed: [Understanding prompt injections: a frontier security challenge]() Original publisher: [Read original article]() Published: 2025-11-07T11:30:00Z Content type: article Language: en Sources: [OpenAI News]() Topics: [prompt injection](), [Security](), [Conversational AI](), [Social engineering](), [Machine Learning, Security Attacks](), [OpenAI]() Tags: [attacks](), [conversational-ai](), [machine-learning](), [openai](), [prompt-injection](), [security]() ### AI overview The article explains prompt injection as a security and social engineering attack against conversational AI. Malicious instructions hidden in webpages, documents, emails, or other third-party content can mislead AI systems into taking unintended actions, especially when they can access sensitive data or act on a user's behalf. It describes OpenAI's focus on recognizing and resisting these attacks through defenses, research, training, and a multi-layered safety approach. ### Source excerpt Prompt injections are a frontier security challenge for AI systems. Learn how these attacks work and how OpenAI is advancing research, training models, and building safeguards for users. ## 4M Models Scanned: Protect AI + Hugging Face 6 Months In DevFeed: [4M Models Scanned: Protect AI + Hugging Face 6 Months In]() Original publisher: [Read original article]() Author: Sean Morgan Published: 2025-04-14T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [hugging face](), [Machine Learning, Security Attacks](), [Threat Research](), [Security](), [threat detection](), [Zero Trust](), [obfuscation](), [vulnerability](), [payload](), [Script](), [llamafile](), [Tensorflow]() Tags: [ai-ml](), [hub](), [hugging-face](), [keras](), [llamafile](), [obfuscation](), [partnerships](), [payload](), [security](), [tensorflow](), [threat-detection](), [vulnerability](), [zero-trust]() ### AI overview Hugging Face and Protect AI report six months of collaboration on Guardian, a scanning system that improves security checks for models hosted on the Hugging Face Hub. Guardian added four detection modules, expanded format and obfuscation coverage, identified a Keras vulnerability, and provides inline alerts and vulnerability reports. The system uses a zero trust approach that treats arbitrary code execution as unsafe, including code hidden through obfuscation. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## Hugging Face and JFrog partner to make AI Security more transparent DevFeed: [Hugging Face and JFrog partner to make AI Security more transparent]() Original publisher: [Read original article]() Author: Luc Georges; Shachar M Published: 2025-03-04T00:00:00Z Content type: news Language: en Sources: [Hugging Face - Blog]() Topics: [Jfrog](), [Securing AI](), [hugging face](), [Exploit](), [Security](), [Machine Learning, Security Attacks](), [Open Source](), [Keras]() Tags: [ai-security](), [exploit](), [hub](), [hugging-face](), [jfrog](), [keras](), [open-source](), [partnerships](), [security]() ### AI overview Hugging Face and JFrog are partnering to improve security on the Hugging Face Hub. JFrog's scanner analyzes code embedded in model weights and supports detection of malicious usage across formats, including pickle and Keras Lambda layers, while public model repositories are scanned automatically. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## Building Trust in a Digital World: The Role of Machine Learning in Behavioral Biometrics DevFeed: [Building Trust in a Digital World: The Role of Machine Learning in Behavioral Biometrics]() Original publisher: [Read original article]() Author: Javier Liébana Published: 2024-06-21T14:01:53Z Content type: article Language: en Sources: [Feedzai]() Topics: [Digital Trust](), [Machine learning](), [Machine Learning, Security Attacks]() Tags: [core](), [digital-trust](), [feedzai](), [fraud-detection](), [fraud-prevention](), [machine-learning](), [online-fraud-prevention](), [research]() ### AI overview This article explains how Feedzai's Digital Trust solution uses behavioral, device, network, operating system, and browser data to support digital identity verification and transaction authorization. It describes a machine learning model that continuously evaluates collected data and uses insights from previous fraud to improve fraud detection and prevention. ### Source excerpt In the world of financial services, the bank or financial institution's relationship with the customer relies on digital trust, which is anchored in two fundamental principles. First, it must ensure the person engaging through digital banking channels is genuinely the individual they claim to be. Second, it must confirm that this person is authorized to complete the intended financial transaction. Addressing these crucial requirements is the core mission of Feedzai's Digital Trust solution. The solution collects and analyzes comprehensive user behavioral data, scrutinizes device information for potential threats, such as malware attacks, and evaluates contextual factors like network, operating system, or browser information to gain a complete understanding of the user's environment. However, the high volume and heterogeneous nature of the collected data, among other challenges, makes detecting potential fraudulent sessions with high accuracy a formidable endeavor. In this blog post we explore how a new machine learning (ML) model that performs a continuous evaluation of collected data and leverages insights from previous frauds to vastly improve Digital Trust's fraud prevention capabilities. We will start with an introduction to technical details behind our Digital Trust solution, going into the challenges of fraud detection and prevention. We continue by explaining how we can apply ML to boost fraud detection and how we deployed the new Fraud model to dozens of Feedzai customers. Table of Contents 1. Digital Trust data collection - 1.1 The user journey - 1.2 Behavioral biometrics data - 1.3 User's behavior - 1.4 Device and network data 2. Challenges to detect fraud in Digital Trust 3. Machine Learning for Digital Trust - 3.1 The holistic approach - 3.2 New Fraud Model 4. Deploying the model 5. In summary Digital Trust data collection To better identify the challenges that are typically faced when designing a fraud prevention system based on Digital Trust, first we ## Nightfall AI and Snyk unite to deliver AI-powered secrets scanning for developers DevFeed: [Nightfall AI and Snyk unite to deliver AI-powered secrets scanning for developers]() Original publisher: [Read original article]() Author: LaToya Muff Published: 2023-11-29T06:00:00Z Content type: article Language: en Sources: [Blog RSS Feed | Snyk]() Topics: [Artificial Intelligence](), [Secret Scanning](), [Security](), [API keys](), [Application Security](), [Cloud](), [Machine Learning, Security Attacks](), [supply-chain-security]() Tags: [acquisition](), [ai](), [api](), [api-keys](), [application-security](), [awareness](), [blog](), [chatgpt](), [devops](), [executive](), [related-content](), [secret-scanning](), [secrets](), [security](), [snyk](), [snyk-platform](), [software-supply-chain](), [supply-chain-security]() ### AI overview Snyk and Nightfall AI partnered to provide AI-powered secrets scanning for developers. Nightfall detects sensitive data such as PII, credentials, secrets, and API keys across cloud environments and supports contextual detection, custom rules, and developer-focused remediation alongside Snyk's code scanning capabilities. ### Source excerpt While Snyk provides a comprehensive approach to developer security by securing critical components of the software supply chain, we recognize the increasing risk of exposed secrets in the cloud. So, we've tapped Nightfall AI to provide a critical feature for developer security: advanced secrets scanning. ## Worthwhile Books Q2 2023 DevFeed: [Worthwhile Books Q2 2023]() Original publisher: [Read original article]() Author: Adam Published: 2023-06-28T00:00:00Z Content type: article Language: en Sources: [Shostack & Friends Blog]() Topics: [Cybersecurity](), [Machine Learning, Security Attacks](), [passwords](), [Zero Trust]() Tags: [attacks](), [books](), [cybersecurity](), [machine-learning](), [password](), [security](), [zero-trust]() ### AI overview A quarterly reading list covering books about cybersecurity, passwords, machine-learning security attacks, zero trust, technology history, and technical writing, along with memoirs and other nonfiction. ### Source excerpt Books that I read in the second quater that are worth your time include two memoirs, a great book on the security of ML, and more! ## Five Threat Model Diagrams for Machine Learning DevFeed: [Five Threat Model Diagrams for Machine Learning]() Original publisher: [Read original article]() Author: Adam Published: 2023-04-13T00:00:00Z Content type: opinion Language: en Sources: [Shostack & Friends Blog]() Topics: [Machine Learning, Security Attacks](), [Machine Learning & Artificial Intelligence](), [Training AI Models](), [Data Quality](), [data]() Tags: [ai-and-cybersecurity](), [cybersecurity](), [data-quality](), [diagram](), [machine-learning](), [training-data]() ### AI overview This article presents five threat-model diagrams for machine learning systems. The diagrams distinguish threats to and from ML systems, illustrate data flows and responses, and examine how training-data sources and system-design decisions can create risks. The author emphasizes that the diagrams are illustrative rather than complete and notes uncertainty about the model of how Twitter content reached Microsoft's Tay. ### Source excerpt Some diagrams to help clarify machine learning threats ## Application Security Roundup - July DevFeed: [Application Security Roundup - July]() Original publisher: [Read original article]() Author: Adam Published: 2022-07-30T00:00:00Z Content type: article Language: en Sources: [Shostack & Friends Blog]() Topics: [Application Security](), [Machine Learning, Security Attacks](), [Architecture & Design]() Tags: [application-security](), [appsec](), [architecture](), [attacks](), [machine-learning](), [performance]() ### AI overview A July roundup of application security posts covering attacks on machine learning systems, Spotify's adaptation of the C4 model for understanding system architecture, and the relationship between performance and security. ### Source excerpt Interesting appsec posts: machine learning, performance, and C4 [Next page]()