Dell SecureWorks’ Post

Interested in learning more about Advanced AI Physical Security?

Security Checklist for Sharing Sensitive Data with AI Tools

Algorithms soon will run your life – and ruin it, if trained incorrectly https://lnkd.in/eMnK-Pbg

Cybercriminals are using generative AI to make sophisticated social engineering attacks. Read how Verizon is creating a custom 5-point plan to help protect your business from the growing social engineering threats. It’s your business. #ItsYourVerizon https://vzbiz.biz/49hXNYy

Sharing latest article for your perusal. https://lnkd.in/gFUmmxXX. This was part of our discussion at DataConnect 2024 Conference. Love sharing research and insights around GenAI for adversarial utilization including Generative Adversarial Networks (GAN) for creating deepfakes. Financial Fraud and reputation risk to your organization of adversarial use of deepfakes should be included in your risk mitigation strategy and Incident Response/Incident Handling planning and tabletop exercises, especially for publicly traded enterprises.

AI can generate a the content for a social engineering attacks in seconds. Do you have a comprehensive Social Engineering Defense plan in place? Take a look at how Verizon is helping our customer's defend against advanced Social Engineering attacks.

Whoops. Building a GPT agent and think you have everything locked down? Scroll to the bottom of the image to see the "security" used to prevent the bot from giving it's instructions to the user.....which it does in the picture. By using prompt injection, we're able to reveal the entire meta-prompt and instructions used to build the custom GPT - even if they tried to instruct the GPT not to do it. Luckily nothing here was terribly damaging nor proprietary, but it's a good lesson on the risks that are at play without proper guidance.

We're now enabling organizations to stay ahead of fraud threats, with readily available data sources that can be easily integrated into decisioning workflows, AI model creation and monitoring. 💪 https://lnkd.in/eWMimS8G

🎉 I'm thrilled to share that I'm back with a brand new blog on "Securing Tomorrow’s AI World Today: Llama Guard Defensive Strategies for LLM Applications." 🏮 🚀 Delving into the crucial realm of AI security risks, this blog is a must-read for anyone navigating the evolving landscape of LLM applications. ✍ ArunVivek Iyer (阿伦维维克) Munirathnakumar Sivaramakrishnan Grateful for the opportunity to explore and contribute to the exciting field of security. Check it out and let's continue the conversation! 🔒🔍 #AIsecurity #LlamaGuard #techblog #datascience