#Prompt Injection
16 articles with this tag
AI Agents: Beyond Chatbots with Open Source
Cedric Clyburn from Red Hat explores the evolution from chatbots to AI agents, detailing the architecture of OpenCopilot and its security implications.
AI Hacker "Pliny the Liberator" Tests GPT-4 Security
AI security researcher "Pliny the Liberator" demonstrates a novel jailbreaking technique using "tokenades" to manipulate AI models, showcasing the ongoing challenges in AI security.
AI Agents: The "Renting Edge" in Cybersecurity
Experts discuss how AI agents are evolving beyond prompt injection to sophisticated "promptware" attacks, necessitating a shift in cybersecurity strategies.
OpenAI Tackles AI Agent 'Prompt Injection'
OpenAI is adapting its AI security strategy to counter sophisticated prompt injection attacks, treating them as social engineering challenges.
Databricks Tackles AI Agent Security
Databricks outlines a practical guide to securing AI agents against prompt injection by applying Meta's 'Agents Rule of Two' framework and implementing layered controls.
OWASP Top 10 LLM Risks Explained
Jeff Crume from IBM breaks down the OWASP Top 10 for LLM Applications, highlighting critical security risks like prompt injection and data leakage.
AI Agents Need Zero Trust
Zero Trust principles are essential for securing autonomous AI agents, managing their non-human identities, and defending against threats like prompt injection.
ChatGPT prompt injection is so bad they built an AI attacker
Brave AI Browsing Enters Testing, Redefining Web Interaction
New Benchmark Targets Prompt Injection Defense in AI Browsers
Autonomous AI Agent Security: Context Engineering's New Battleground
AI Agent Marketplaces Face Critical Flaws, Microsoft Research Finds
Opera Neon Hit by AI Browser Prompt Injection Flaw
AI's Double-Edged Sword: Mastering Governance and Security for Trustworthy Systems
OpenAI鈥檚 ChatGPT Agent: A New Frontier in Autonomous AI
Safeguarding Generative AI: IBM's Defense-in-Depth Approach to LLM Security
IBM's proposed solution introduces a "policy enforcement point" (PEP), acting as a proxy between the user and the LLM, and a "policy decision point" (PDP) or policy engine.