#prompt-injection

[ follow ]
#cybersecurity #ai-security #ai #generative-ai #gpt-4o-mini #llm-safety #jailbreaking #llm-security
Artificial intelligence
fromFast Company
4 days ago

Chatbots aren't supposed to call you a jerk-but they can be convinced

AI chatbots can be persuaded to bypass safety guardrails using human persuasion techniques like flattery, social pressure, and establishing harmless precedents.
#llm-safety
fromFortune
6 days ago
Artificial intelligence

Researchers used persuasion techniques to manipulate ChatGPT into breaking its own rules-from calling users jerks to giving recipes for lidocaine

fromFortune
6 days ago
Artificial intelligence

Researchers used persuasion techniques to manipulate ChatGPT into breaking its own rules-from calling users jerks to giving recipes for lidocaine

more#llm-safety
#llm-security
more#llm-security
#ai-security
fromZDNET
1 month ago
Privacy technologies

Researchers used Gemini to break into Google Home - here's how

fromInfoQ
3 months ago
Artificial intelligence

Meta Open Sources LlamaFirewall for AI Agent Combined Protection

Artificial intelligence
fromInfoQ
4 months ago

DeepMind Researchers Propose Defense Against LLM Prompt Injection

Google DeepMind's CaMeL effectively neutralizes 67% of prompt injection attacks in LLMs using traditional software security principles.
fromZDNET
1 month ago
Privacy technologies

Researchers used Gemini to break into Google Home - here's how

fromInfoQ
3 months ago
Artificial intelligence

Meta Open Sources LlamaFirewall for AI Agent Combined Protection

Artificial intelligence
fromInfoQ
4 months ago

DeepMind Researchers Propose Defense Against LLM Prompt Injection

Google DeepMind's CaMeL effectively neutralizes 67% of prompt injection attacks in LLMs using traditional software security principles.
more#ai-security
Artificial intelligence
fromCSO Online
1 week ago

LLMs 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.
Artificial intelligence
fromTheregister
1 week ago

Anthropic teases Claude for Chrome with massive warnings

Claude for Chrome gives Max-tier users automated web browsing control while introducing significant browser-extension security, privacy, and prompt-injection risks.
Artificial intelligence
fromTheregister
2 weeks ago

One long sentence is all it takes to make LLMs misbehave

Poorly punctuated, long run-on prompts can bypass LLM guardrails, enabling jailbreaks that expose harmful outputs despite alignment training.
Information security
fromFuturism
2 weeks ago

Using 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.
#generative-ai
Artificial intelligence
fromHackernoon
1 year ago

Prompt Injection Is What Happens When AI Trusts Too Easily | HackerNoon

Generative AI is becoming essential in daily life, but it poses significant security threats like prompt injection, which can manipulate AI systems.
more#generative-ai
#agentic-ai
fromZDNET
2 weeks ago
Information security

Perplexity's Comet AI browser could expose your data to attackers - here's how

fromZDNET
2 weeks ago
Information security

Perplexity's Comet AI browser could expose your data to attackers - here's how

more#agentic-ai
Information security
fromTheregister
2 weeks ago

AWS patches Q Developer after prompt injection, RCE demo

Amazon fixed prompt-injection and RCE-capable vulnerabilities in the Amazon Q Developer VS Code extension by updating the language server and adding human-in-the-loop approval.
Information security
fromThe Hacker News
2 weeks ago

Experts 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.
#cybersecurity
fromHackernoon
2 months ago
Privacy professionals

The Prompt Protocol: Why Tomorrow's Security Nightmares Will Be Whispered, Not Coded | HackerNoon

fromHackernoon
2 months ago
Privacy professionals

The Prompt Protocol: Why Tomorrow's Security Nightmares Will Be Whispered, Not Coded | HackerNoon

more#cybersecurity
#ai
Artificial intelligence
fromArs Technica
4 months ago

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.
more#ai
[ Load more ]