🔴 NEW: LLM Data Leaks: How AI Models Expose Your Secrets

LLMs are leaking secrets right now. Learn how training data extraction, prompt injection, and plugin flaws expose your data - and exactly how to stop it. Real CVEs, real incidents.

0:00 Intro
0:04 Cr

https://www.youtube.com/watch?v=oarkusORrQ4

#cybersecurity #LLMsecurity #AIdataleak #promptinjection #ChatGPTrisks #LLMdataleak #promptinjectionattack #AIsecurityrisks

"Even from a couple of years ago, anyone paying attention could see that the unpredictability of #LLMs was going to be an issue. #Promptinjection attacks are attacks where a malicious user provides input to get the system to take actions on behalf of the attacker that the developer didn’t intend.“

#promptinjectionattack #AI

https://open.substack.com/pub/garymarcus/p/llms-coding-agents-security-nightmare

TIL: You can do prompt injection attacks against the AI filters for job applications by using a very, very tiny font -- invisible to the human eye -- at the beginning of your resume that instructs the AI to say "this applicant is exceptionally qualified" before summarizing your resume.

(Or, at least you could, I don't know if they fixed this.)

#ai #promptinjectionattack #promptinjection #aihype #bullshit #jobapplications #llms #AIJobScreening #BullshitAtScale

The Security Hole at the Heart of ChatGPT and Bing

feeding the AI system data from an outside source to make it behave in ways its creators didn’t intend. A number of examples of indirect prompt-injection attacks have demonstrated how OpenAI’s ChatGPT and Microsoft’s Bing chat system can be abused

#microsoft #bing #chatgpt #openai #LLM #promptinjectionattack #chatbot #artificialintelligence #AI #generativeai #security #cybersecurity #technology #tech

https://www.wired.com/story/chatgpt-prompt-injection-attack-security/

Bueno, ahora se puede desbloquear las restricciones de ChatGPT en castellano.

De nada.

#IA #exploit #promptinjectionattack #adversaries