Explore tweets tagged as #PromptInjection
๐๐ซ๐จ๐ฆ๐ฉ๐ญ ๐๐ง๐ฃ๐๐๐ญ๐ข๐จ๐ง ๐๐ญ๐ญ๐๐๐ค ๐ฅ๐๐๐ฌ ๐ฐ๐ข๐ญ๐ก ๐ซ๐๐๐ฅ-๐ญ๐ข๐ฆ๐ ๐๐๐๐๐๐๐๐ค ๐๐ง๐ ๐ซ๐๐๐ข๐ง๐ ๐ฒ๐จ๐ฎ๐ซ ๐๐ ๐ข๐ง๐ญ๐๐ซ๐๐๐ญ๐ข๐จ๐ง๐ฌ ๐จ๐ง ๐ญ๐ก๐ ๐ ๐จ. Check it out here: ๐ #AI #CyberSecurity #AISecurity #PromptInjection
1
3
8
[New on our blog] Understanding #PromptInjection: Risks, Methods, and Defense Measures . Author: Soumya Shaw.Reading time: 13 min. โ. Full article:
0
0
0
Mitigating prompt-injection in software agents โผ๏ธ. The Good ๐.The Bad ๐.And a playbook on how to stay safe ๐ฆพ. Read to learn more๐. #AIsecurity #LLMAgents #PromptInjection
2
3
23
I've managed to get GPT-4 to predict my future - and believe it or not, it even came true! ๐๐. What's interesting about this File Name #PromptInjection technique is its delayed trigger, allowing activation at any stage of the conversation. Nice find @elder_plinius!
0
1
2
Guardrails, Prompt Injection, Evaluation, Human-in-the-Loop โ essential safeguards for deploying GenAI systems.#Guardrails #PromptInjection #Evaluation #HumanInTheLoop #HITL #AItrust #AISafety #AIEvaluation #LLM #promptengineering #systemprompt #outputfiltering #biasdetection
2
0
2
๐ Subscribe to my newsletter and read the details article - ๐จ THE AI HACK THAT'S BREAKING THE INTERNET: The Prompt Injection Pandemic. ๐ > #PromptInjection #AIHack #OpenAI #ChatGPT #CyberSecurity #DataBreach #AIThreats #PromptInjection #AIHack #ChatGPT
1
0
1
With the new ChatGPT Search feature, Indirect Prompt Injection occurs when you directly visit the URL. It also wasnโt visible in the sources. ๐ฝ. #promptinjection #chatgpt
1
1
10
El lado del mal - Hacking IA: Indirect Prompt Injection en Perplexity Comet #PromptInjection #Perplexity #Comet #IA #AI #Hacking #Pentest #Pentesting #InteligenciaArtificial #AgenticAI
3
136
158
What is Prompt Injection? Risks, Examples, and How to Prevent Attacks. #ai #artificialintelligence #prompt #promptengineering #promptinjection #hacker.
0
0
0
๐ฅ How Hackers Weaponize Language ๐ง .#AIHacking #PromptInjection #CyberAwareness #AIChatbots #DataSecurity #Infosec.๐ Subscribe to my newsletter and read the details article - ๐จ THE AI HACK THAT'S BREAKING THE INTERNET: The Prompt Injection Pandemic.๐ >
2
0
2
๐คฆ๐ปโโ๏ธ Oh great, another new AI attack surface. prompt injection hiding in images. Looks harmless, but resized images can trigger hidden commands & even exfiltrate data. ๐ Needs layered defenses. #AIsecurity #promptinjection #cybersecurity.
2
0
0
๐ค๐ฅ Prompt injection is still one of the most underestimated risks in AI development. If you're building with or around LLMs, this is worth a read:. ๐ #AIsecurity #PromptInjection #AppSec #LLMSecurity #CyberSecurity #AIrisks
0
0
3
AI tools are useful, but still need control. A recent AWS Copilot incident revealed "malicious" attempts at supply chain attacks, leading to disruptive actions. #AIsecurity #PromptInjection
0
0
1
Posted on Upwork and one of my questions used prompt injection to only have LLMs reply as a pirate .#upwork #promptinjection
0
0
0
El lado del mal - PROMISQROUTE para GPT-5: Un ataque de downgrade forzando el Routing para hacer Jailbreak #IA #AI #Jailbreak #PromptInjection #ChatGPT #GPT #OpenAI #bug #Hacking #Pentest #Pentesting #InteligenciaArtificial
3
135
168
No model is immune. Even ChatGPT-4o and Copilot can be steered under the right conditions. Read our findings:.#LLMSecurity #PromptInjection #AIExploitation #CyberThreats.
0
0
1
Jason Haddixโs AI Attack Taxonomy.#JasonHaddix #AIHack #ArcanumInfoSec #PromptInjection #CyberSecurity #ArcanumInfoSec #Infosec.๐ Subscribe to my newsletter and read the details article - ๐จ THE AI HACK THAT'S BREAKING THE INTERNET: The Prompt Injection Pandemic.๐ >
0
0
1