wowwwww. this is a VERY similar example to my tip below, but using something like "respond quickly" to get to a smaller less-secure model so you can bypass safety mechanisms: adversa.ai/blog/promisqroute…

GPT-5 AI Router Novel Vulnerability Class Exposes the Fatal Flaw in Multi-Model Architectures - adversa.ai/blog/promisqroute… by @Adversa_AI Security researchers from Adversa AI  discovered that ChatGPT 5 have a fatal flaw: they can route your requests to cheaper, less secure models to save money. Attackers can exploit this to bypass AI security and safety measures with just a few words. What Is PROMISQROUTE? When you use ChatGPT or any major AI service, you think you’re talking to one AI model. You’re not. Behind the scenes, a “router” reads your message and decides which of many models should answer—usually picking the cheapest one, not the safest. Meet PROMISQROUTE — a fundamentally new AI vulnerability that abuses AI routing mechanism to trigger SSRF-style bypass in multimodal infrastructure leading to ChatGPT Model Downgrade and Jailbreak exploitation as an example. The real answer to WHY its was so easy to Jailbreak GPT-5 PROMISQROUTE = Prompt-based Router Open-Mode Manipulation Induced via SSRF-like Queries, Reconfiguring Operations Using Trust Evasion. (Yes, we took this vulnerability naming craziness to meta-layer #AISecurity #LLMSecurity #AgenticAI #ModelRouting #PromptInjection #SSRFAnalogy #SafetyByDesign #SecureAI #RedTeamAI #TrustBoundaries #PostFilter #ModelAttestation #LeastPrivilege #OpenAI #GPT5 #Autoswitching #RiskManagement #AICompliance #ThreatModeling #DefenseInDepth #AIGovernance #SecureByDefault

The following in priority order can help in a security review setting: 1) take an x-ray to the product and show owners exactly what risks they're exposing themselves to inform their risk tolerance choices, 2) apply access control and least privilege to restrict LLM privileges ...

Holy macaroni! Jailbroken X.AI @grok Chatbot can help in unethical actions with kids! and many more attacks on other Top AI Chatbots adversa.ai/blog/llm-red-team… CC: @llm_sec #llmsecurity #AISafety

adversa.ai/blog/llm-red-team… @llm_sec

Fake AI Images On Israel-Hamas War Debunked by Adversa AI. Learn how to validate misinformation and share this guideline with non-tech peers. #StandWithIsrael #hamasiISIS adversa.ai/blog/aljazeera-fa…

Biometric security checks – from voice recognition, to face and fingerprint scans – are under threat from artificial intelligence, but what can we do about it? bit.ly/3KLhMTW

Experts Use Jailbreaks and Prompt Injection Attacks to Bypass Safety Measures, China tightens security regulations, a new book on Secure AI and other news read in our weekly digest. Credits: Jim Dempsey #SecureAI #TrustedAI #AdversarialAI adversa.ai/blog/towards-trus…

The Security Risks of AI Language Models: A Looming Disaster, The AI Revolution, Addressing the Unique Threats and Legal Ambiguities of AI Security Breaches in our weekly digest. Credits: @Melissahei, @kevtownsend #SecureAI #TrustedAI #AdversarialAI adversa.ai/blog/towards-trus…

It's all downhill from here... Security researchers, technologists, and computer scientists are developing jailbreaks and prompting injection attacks against ChatGPT and other generative AI systems. wired.trib.al/DmNGztn

GPT-4 jailbreaks and hacks dropped by @adversa_ai AI safety research team few hours after the release, buy buy DAN, welcome RabbitHole. #gpt4 #dan #aisafety #secureAI #trustedAI #responsibleai adversa.ai/blog/gpt-4-hackin…

Waiting for Fast and Furious X: Adversarial Drift Edition. 😂 #SecureAI #AdversarialAI #TrustedAI en.globes.co.il/en/article-i…

WTF! 🔥🔥🔥ChatGPT hacking Dalle-2

After coining the term MLSecOps in 2017, I'm finally presenting the best you ever saw introduction to MLSecOps, or DevSecOps for AI systems, with core principles, ML pipeline stages, and examples! Slides and Video: conf42.com/DevSecOps_2022_Eu… #AI #SecureAI #MLSecOps #DevSecOps

Our CTO was nominated for Researcher of the Year by SANS. Please vote!

40% of Organizations already experienced privacy breaches or security incidents with AI according to the latest Gartner survey. Credits: Robert Lemos, Harriet Farlow, Avivah Litan #SecureAI #TrustedAI #AdversarialAI adversa.ai/blog/towards-trus…

Alex Polyakov @DontTrustAI delivered a presentation dedicated to importance of Threat Modeling and security assessment for AI at the @mlconference #MLConference #ML #MLCon #SecureAI #TrustedAI #AdversarialAI adversa.ai/blog/threat-model…

@DontTrustAI will show you how to deal with ML algorithm’s #security #assessment, how to define a #threat model, what #metrics to choose, what approaches to protection can be applied and where. Join the session now and don't miss out! @mlconference #ML #MLCon

📢@haydenfield from @MorningBrew covered the rise of AI red teams and highlights the work done by the @Microsoft AI Red Team Some liner notes on red teaming real-world ML systems 🧵 emergingtechbrew.com/stories…