Tag: chatgpt
-
Widely Used Malicious Extensions Steal ChatGPT, DeepSeek Conversations
Threat actors used two malicious Chrome extensions that have 900,000 users to steal their chats with AI models like ChatGPT and DeepSeek and browser history. The incident is the latest in a growing string of attacks in which hackers weaponized browser extensions to exfiltrate huge amounts of sensitive data. First seen on securityboulevard.com Jump to…
-
OpenAI says prompt injection may never be ‘solved’ for browser agents like Atlas
OpenAI is warning that prompt injection, a technique that hides malicious instructions inside ordinary online content, is becoming a central security risk for AI agents designed to operate inside a web browser and carry out tasks for users. The company said it recently shipped a security update for ChatGPT Atlas after internal automated red-teaming uncovered…
-
Top 5 real-world AI security threats revealed in 2025
Tags: access, ai, api, attack, breach, chatgpt, cloud, control, credentials, cybercrime, data, data-breach, defense, email, exploit, flaw, framework, github, gitlab, google, injection, least-privilege, LLM, malicious, malware, microsoft, nvidia, open-source, openai, rce, remote-code-execution, risk, service, software, supply-chain, theft, threat, tool, vulnerabilityA critical remote code execution (RCE) in open-source AI agent framework Langflow that was also exploited in the wildAn RCE flaw in OpenAI’s Codex CLIVulnerabilities in NVIDIA Triton Inference ServerRCE vulnerabilities in major AI inference server frameworks, including those from Meta, Nvidia, Microsoft, and open-source projects such as vLLM and SGLangVulnerabilities in open-source compute framework…
-
OpenAI’s ChatGPT ads will allegedly prioritize sponsored content in answers
OpenAI is reportedly mulling a new form of ads on ChatGPT called “sponsored content,” which could influence your buying decisions. First seen on bleepingcomputer.com Jump to article: www.bleepingcomputer.com/news/artificial-intelligence/openais-chatgpt-ads-will-allegedly-prioritize-sponsored-content-in-answers/
-
OpenAI’s ChatGPT ads will allegedly prioritize sponsored content in answers
OpenAI is reportedly mulling a new form of ads on ChatGPT called “sponsored content,” which could influence your buying decisions. First seen on bleepingcomputer.com Jump to article: www.bleepingcomputer.com/news/artificial-intelligence/openais-chatgpt-ads-will-allegedly-prioritize-sponsored-content-in-answers/
-
OpenAI is reportedly testing Claude-like Skills for ChatGPT
OpenAI is testing a new ChatGPT feature called “Skills,” which will be similar to Claude’s feature, also called Skills. First seen on bleepingcomputer.com Jump to article: www.bleepingcomputer.com/news/artificial-intelligence/openai-is-reportedly-testing-claude-like-skills-for-chatgpt/
-
OpenAI Will Forever Fight Prompt Injection Attacks
AI Firm Discovers New Prompt Injection Attack Class. OpenAI faces a years-long battle to secure its ChatGPT Atlas web browser against prompt injection attacks, a threat the company says will require continuous defense strengthening much like the arms race against online scams targeting humans. First seen on govinfosecurity.com Jump to article: www.govinfosecurity.com/openai-will-forever-fight-prompt-injection-attacks-a-30380
-
Von 6 Millionen Nutzern installiert: Diese Chrome-Erweiterung sammelt laut Experten heimlich deine ChatGPT-Daten
First seen on t3n.de Jump to article: t3n.de/news/6-millionen-nutzer-installiert-chrome-erweiterung-sammelt-heimlich-chatgpt-daten-1721943/
-
Von 6 Millionen Nutzern installiert: Diese Chrome-Erweiterung sammelt laut Experten heimlich deine ChatGPT-Daten
First seen on t3n.de Jump to article: t3n.de/news/6-millionen-nutzer-installiert-chrome-erweiterung-sammelt-heimlich-chatgpt-daten-1721943/
-
Staatsgeheimnis: KI-Nutzung bei Patentanmeldung hebt Geheimhaltung auf
Eine Patentanmeldung wird erst als Staatsgeheimnis eingestuft, dann wieder freigegeben. Der Grund: Der Erfinder hatte ChatGPT genutzt. First seen on golem.de Jump to article: www.golem.de/news/staatsgeheimnis-ki-nutzung-bei-patentanmeldung-hebt-geheimhaltung-auf-2512-203387.html
-
NDSS 2025 IsolateGPT: An Execution Isolation Architecture For LLM-Based Agentic Systems
Session 6A: LLM Privacy and Usable Privacy Authors, Creators & Presenters: Yuhao Wu (Washington University in St. Louis), Franziska Roesner (University of Washington), Tadayoshi Kohno (University of Washington), Ning Zhang (Washington University in St. Louis), Umar Iqbal (Washington University in St. Louis) PAPER IsolateGPT: An Execution Isolation Architecture for LLM-Based Agentic Systems Large language models…
-
Browser Extension Harvests 8M Users’ AI Chatbot Data
Urban VPN Proxy, which claims to protect users’ privacy, collects data from conversations with ChatGPT, Claude, Gemini, Copilot and other AI assistants. First seen on darkreading.com Jump to article: www.darkreading.com/endpoint-security/chrome-extension-harvests-ai-chatbot-data
-
Chrome Extension with 6M+ Users Found Collecting AI Chatbot Inputs
A popular browser extension promoted as a free and secure VPN has been discovered secretly capturing user conversations across multiple AI chatbot platforms including ChatGPT, Claude, Gemini, and Microsoft Copilot raising fresh concerns over privacy and data exploitation in the age of generative AI. Researchers using the Wings agentic”‘AI risk engine uncovered that Urban VPN…
-
Featured Chrome Browser Extension Caught Intercepting Millions of Users’ AI Chats
A Google Chrome extension with a “Featured” badge and six million users has been observed silently gathering every prompt entered by users into artificial intelligence (AI)-powered chatbots like OpenAI ChatGPT, Anthropic Claude, Microsoft Copilot, DeepSeek, Google Gemini, xAI Grok, Meta AI, and Perplexity.The extension in question is Urban VPN Proxy, which has a 4.7 rating…
-
Security for AI: How Shadow AI, Platform Risks, and Data Leakage Leave Your Organization Exposed
Tags: access, ai, attack, awareness, business, chatgpt, china, cloud, compliance, control, corporate, cybersecurity, data, data-breach, defense, detection, endpoint, governance, guide, infrastructure, injection, leak, LLM, malicious, microsoft, mitigation, monitoring, network, open-source, openai, privacy, RedTeam, risk, saas, service, strategy, threat, tool, training, vulnerabilityYour employees are using AI whether you’ve sanctioned it or not. And even if you’ve carefully vetted and approved an enterprise-grade AI platform, you’re still at risk of attacks and data leakage. Key takeaways: Security teams face three key risks as AI usage becomes widespread at work: Shadow AI, the challenge of safely sanctioning tools,…
-
Security for AI: How Shadow AI, Platform Risks, and Data Leakage Leave Your Organization Exposed
Tags: access, ai, attack, awareness, business, chatgpt, china, cloud, compliance, control, corporate, cybersecurity, data, data-breach, defense, detection, endpoint, governance, guide, infrastructure, injection, leak, LLM, malicious, microsoft, mitigation, monitoring, network, open-source, openai, privacy, RedTeam, risk, saas, service, strategy, threat, tool, training, vulnerabilityYour employees are using AI whether you’ve sanctioned it or not. And even if you’ve carefully vetted and approved an enterprise-grade AI platform, you’re still at risk of attacks and data leakage. Key takeaways: Security teams face three key risks as AI usage becomes widespread at work: Shadow AI, the challenge of safely sanctioning tools,…
-
Hackers Are Using Shared AI Chats to Steal Your Passwords and Crypto
A sophisticated malvertising campaign is exploiting ChatGPT and DeepSeek’s shared chat features to deliver credential-stealing malware to macOS users. Threat actors are purchasing sponsored Google search results and redirecting victims to legitimate-looking LLM-generated chat sessions that contain obfuscated malicious commands, effectively bypassing platform-level safety mechanisms. The attack begins when users search for common macOS troubleshooting…
-
Hackers Are Using Shared AI Chats to Steal Your Passwords and Crypto
A sophisticated malvertising campaign is exploiting ChatGPT and DeepSeek’s shared chat features to deliver credential-stealing malware to macOS users. Threat actors are purchasing sponsored Google search results and redirecting victims to legitimate-looking LLM-generated chat sessions that contain obfuscated malicious commands, effectively bypassing platform-level safety mechanisms. The attack begins when users search for common macOS troubleshooting…
-
Google ads for shared ChatGPT, Grok guides push macOS infostealer malware
A new AMOS infostealer campaign is abusing Google search ads to lure users into Grok and ChatGPT conversations that appear to offer “helpful” instructions but ultimately lead to installing the AMOS info-stealing malware on macOS. First seen on bleepingcomputer.com Jump to article: www.bleepingcomputer.com/news/security/google-ads-for-shared-chatgpt-grok-guides-push-macos-infostealer-malware/
-
KI-Browser gefährden Unternehmen
Experten warnen vor der Nutzung von KI-Browsern in Unternehmen.Die Gartner-Analysten Dennis Xu, Evgeny Mirolyubov und John Watts empfehlen Unternehmen dringend, alle KI-Browser aufgrund der Cybersicherheitsrisiken auf absehbare Zeit zu blockieren. Sie stützten ihre Empfehlung auf bereits identifizierte Risiken ‘und andere potenzielle Risiken, die noch entdeckt werden müssen, da es sich um eine sehr junge Technologie…
-
KI-Browser gefährden Unternehmen
Experten warnen vor der Nutzung von KI-Browsern in Unternehmen.Die Gartner-Analysten Dennis Xu, Evgeny Mirolyubov und John Watts empfehlen Unternehmen dringend, alle KI-Browser aufgrund der Cybersicherheitsrisiken auf absehbare Zeit zu blockieren. Sie stützten ihre Empfehlung auf bereits identifizierte Risiken ‘und andere potenzielle Risiken, die noch entdeckt werden müssen, da es sich um eine sehr junge Technologie…
-
Gemini for Chrome gets a second AI agent to watch over it
Google’s two-model defense: To address these risks, Google’s solution splits the work between two AI models. The main Gemini model reads web content and decides what actions to take. The user alignment critic sees only metadata about proposed actions, not the web content that might contain malicious instructions.”This component is architected to see only metadata…