Yonah Welker’s Post

As our Eviden Global CTO digital security Zeina Zakhour, says "Solving the #AI + #cybersecurity equation is not just any mathematical challenge. It is a multi-variable puzzle where AI and cybersecurity intersect, overlap, and diverge. If X equals AI and Y equals cybersecurity, Z is the unknown in this cubic equation with variables that refuse to stay constant: new technologies, new tactics by cybercriminals, or the ethical implications of AI decisions..." Solve the equation in our Digital Security Magazine https://lnkd.in/e_ATThHH #eviden #cyberthreats #ethicalai #artificialintelligence https://lnkd.in/etGFV_xC

𝐀𝐈 𝐆𝐨𝐧𝐞 𝐖𝐢𝐥𝐝: 𝐓𝐡𝐞 𝐑𝐢𝐬𝐤𝐬 𝐨𝐟 𝐄𝐱𝐜𝐞𝐬𝐬𝐢𝐯𝐞 𝐀𝐮𝐭𝐨𝐧𝐨𝐦𝐲 𝐒𝐮𝐦𝐦𝐚𝐫𝐲: This article explores the risks of granting large language models (LLMs) excessive autonomy. When LLMs are given too much control over sensitive systems, they can make mistakes, cause unintended consequences, and create security vulnerabilities. 𝐊𝐞𝐲 𝐏𝐨𝐢𝐧𝐭𝐬: ♦ Excessive agency can lead to erroneous decisions and unintended consequences. ♦ Mitigating excessive agency requires limiting autonomy, implementing oversight mechanisms, and using role-based access control. ♦ Case studies highlight the real-world risks of excessive agency, such as LLM-induced stock market manipulation and autonomous document publishing without review. https://lnkd.in/enCvHYv5 #LLM #AI #ArtificialIntelligence #MachineLearning #AIethics #DataPrivacy #Cybersecurity

Over the last several months, I have been increasingly impressed with the innovative solutions my colleagues in #datascience and #cybersecurity have been creating for the market. So, when I heard they were leveraging #AI and #LLMs to help create efficiencies and drive more effective outcomes for our clients, I knew we needed to dig deeper. Read more in our case study below on how this new technology can help analyze large volumes of disparate data, and where we see it going next.   #artificialintelligence #dataanalytics 

In today’s landscape, the convergence of Artificial Intelligence (AI) and Cybersecurity has become more crucial than ever. In this video, Andreas Fuchsberger, Chair of IST/33/3 (Security Evaluation, Testing, and Specification), explores how these two powerful forces interact and what they mean for your organization. Dive deep into cutting-edge strategies and standards that help organizations safeguard their data while leveraging the potential of AI. At BSI, we're committed to leading the conversation on securely navigating this new frontier. Watch the video to discover how our expertise in cybersecurity standards can empower your business to stay ahead. Watch the video: https://bit.ly/48t5iMg #Cybersecurity #AI #BSIStandards

NIST’s latest insights emphasize the unique cybersecurity and privacy challenges posed by artificial intelligence. Key strategies include: Risk Assessment: Identify specific AI-related risks, including data privacy issues. Security by Design: Integrate security measures during AI development to minimize vulnerabilities. Strong Governance: Establish frameworks for accountability and regulatory compliance. Ongoing Monitoring: Implement continuous risk monitoring to stay ahead of evolving threats. Stakeholder Engagement: Foster collaboration across teams to enhance security awareness. As AI technology grows, prioritizing cybersecurity and privacy is essential for safeguarding organizations and users alike! It's nice to see NIST react quickly with regards to AI. #Cybersecurity #AI #Privacy #NIST #RiskManagement https://lnkd.in/eUYJBq-5

Want to glean insight into AI issues impacting corporations? My colleague, Kim Peretti, will moderate the “Cybersecurity and AI – Challenges, Opportunities, and Trends” panel on Oct 8. Thought leaders will explore some of the following topics:    - Challenges and opportunities of #AI  - How to prepare for a changing threat landscape  - Best practices for companies to reduce #Cybersecurity risks    Register for the AI Legal Insights: Shaping Tomorrow – Program 2 webinar.  https://bit.ly/4eL9by3

Bring some fun to work. “Your LLM might have less discretion than you’d like when handling sensitive information. This vulnerability, which the OWASP Top 10 for LLMs calls “sensitive information disclosure,” occurs when the model inadvertently outputs confidential or sensitive data it has gleaned from its extensive training, such as passwords or personal details.” From: The Developer's Playbook for Large Language Model Security: Building Secure AI Applications, O'Reilly Media. Author: Steve Wilson. #AI #ArtificialIntelligence #MachineLearning #AIsecurity #Cybersecurity #DataPrivacy #DataSecurity #InformationSecurity #LLM #LargeLanguageModel #GenerativeAI #PromptInjection #AIethics #AIethics

There has been great speculation of how #AI will transform business, but how does this technology apply in practice? StoneTurn’s #cybersecurity, data science, and #risk and controls experts discuss how they developed an LLM-based review to cross-reference and assess large volumes of information stored in documents and files in different formats against the National Institute of Standards and Technology’s Cybersecurity Framework. Read more from Ian Costa, Michael Costa, Kashif Sheikh, JD, CFA, Luke Tenery and Jonny Frank on how this technology can catalyze processes and analysis: https://lnkd.in/ekYXW6Vf #LLM #largelanguagemodel #artificialintelligence #datascience #analytics #compliance

Happy Monday! As you set up time for podcasts to listen to during commutes, workouts, and walks this week, add Advancing Cyber and "Should AI be Critical Infrastructure?" to the list! Jason Healey asks whether it's time for an AI ISAC, Jesper Johansson reminds us that AI isn't meant to be deterministic, and asks how do we even know who the victims will be? David Simon reminds us that companies have to think about which regulators are "in play" for AI and incidents, and that the list will grow. Listen now on Spotify: https://lnkd.in/gwSBhQar Listen now on Apple Podcast: https://lnkd.in/g5bqX7iu #ArtificialIntelligence #Cybersecurity #IncidentResponse #ISACs Advancing Cyber Advanced Cyber Law #AdvancingCyber

Congress is pushing boundaries and lighting fires with the AI Incident Reporting and Security Enhancement Act 🔥 This bold move attacks the volatile intersection of AI and cybersecurity, pushing the National Institute of Standards and Technology [NIST] to include AI vulnerabilities in the National Vulnerability Database [NVD]. Double the chaos? No, double the protection! 🟠 NIST's new mandate to track AI system vulnerabilities will shake up how we report and mitigate threats. 🟠 A standardized process will make it crystal clear for organizations to pinpoint and smash potential risks. 🟠 This isn't just a tweak. It's a groundbreaking effort to solidify AI's defense systems—acknowledging the mounting risks that come with soaring advances. Are we ready to face the dark side of AI head-on or are we opening Pandora's box? 🚀 Drop your thoughts below! 👇 #Cybersecurity #AI #AIVulnerabilities #NIST #TechTrends