#prompt-injection
#prompt-injection
[ follow ]
#ai-security #data-exfiltration #cybersecurity #ai #google-gemini #agentic-ai #llm-security #salesforce-agentforce
Information security
fromThe Hacker News
2 days agoResearchers Disclose Google Gemini AI Flaws Allowing Prompt Injection and Cloud Exploits
Three now-patched vulnerabilities in Google's Gemini AI could have allowed prompt and search-injection attacks and user data exfiltration across cloud services and browsing tools.
fromThe Hacker News
1 week agoSalesforce Patches Critical ForcedLeak Bug Exposing CRM Data via AI Prompt Injection
Cybersecurity researchers have disclosed a critical flaw impacting Salesforce Agentforce, a platform for building artificial intelligence (AI) agents, that could allow attackers to potentially exfiltrate sensitive data from its customer relationship management (CRM) tool by means of an indirect prompt injection. The vulnerability has been codenamed ForcedLeak (CVSS score: 9.4) by Noma Security, which discovered and reported the problem on July 28, 2025. It impacts any organization using Salesforce Agentforce with the Web-to-Lead functionality enabled.
Information security
fromZDNET
1 week agoHow researchers tricked ChatGPT into sharing sensitive email data
While AI agents show promise in bringing AI assistance to the next level by carrying out tasks for users, that autonomy also unleashes a whole new set of risks. Cybersecurity company Radware, as by The Verge, decided to test OpenAI's Deep Research agent for those risks -- and the results were alarming. Also: OpenAI's Deep Research has more fact-finding stamina than you, but it's still wrong half the time
Information security
fromSecurityWeek
1 week agoChatGPT Tricked Into Solving CAPTCHAs
AI agents have guardrails in place to prevent them from solving any CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart), based on ethical, legal, and platform-policy reasons. When asked directly, a ChatGPT agent refuses to solve a CAPTCHA, but anyone can apparently use misdirection to trick the agent into giving its consent to solve the test, and this is what SPLX demonstrated.
Artificial intelligence
Artificial intelligence
fromBusiness Insider
2 weeks agoThe billion-dollar rivalry over 'ChatGPT for doctors' just got nastier with dueling lawsuits
Two major healthcare AI companies, Doximity and OpenEvidence, are engaged in escalating lawsuits over alleged impersonation, prompt-injection attacks, trade-secret theft, and reputational harm.
Artificial intelligence
fromCSO Online
1 month agoLLMs easily exploited using run-on sentences, bad grammar, image scaling
Large language models remain easily manipulated into revealing sensitive data via prompt formatting and hidden-image attacks due to alignment training gaps and brittle prompt security.
Information security
fromFuturism
1 month agoUsing an AI Browser Lets Hackers Drain Your Bank Account Just by Showing You a Public Reddit Post
Perplexity's Comet browser AI accepts webpage content as commands, enabling simple indirect prompt injections that can grant attackers access to user accounts and private data.
Information security
fromThe Hacker News
1 month agoExperts Find AI Browsers Can Be Tricked by PromptFix Exploit to Run Malicious Hidden Prompts
PromptFix hides malicious instructions inside fake CAPTCHA checks to trick GenAI browsers and agentic AI into interacting with phishing sites and performing attacker actions.
#ai
fromArs Technica
5 months agoArtificial intelligence
Researchers claim breakthrough in fight against AI's frustrating security hole
Prompt injections jeopardize AI systems; Google DeepMind's CaMeL offers a potential solution by treating language models as untrusted components within security frameworks.
Artificial intelligence
fromArs Technica
5 months agoResearchers claim breakthrough in fight against AI's frustrating security hole
Prompt injections jeopardize AI systems; Google DeepMind's CaMeL offers a potential solution by treating language models as untrusted components within security frameworks.
[ Load more ]