OWASP AI Exchange

Our very own Susanna spreading the AI security word.

Watch the recording of our very own Eng. Behnaz Karimi giving a nice overview of the AI Exchange at the OWASP Germany day 2024. Well done Behnaz! #ai #aisecurity

The OWASP AI Exchange team regularly posts highlights from the material at owaspai.org. This week it's the References section, that today received a massive extension with training material, CTF resources, and AI security talks - kudos to author Zia Rashid. Have a look at https://lnkd.in/eeVSbXjW You may want to bookmark it. In fact, that is an important goal for us: be a primary bookmark for AI security practitioners out there. #ai #aisecurity

Isn't it amazing and surreal, to finally meet people in person after years of working together online? The OWASP? Foundation is reporting live from Torino Italy, from its seat at the international standardization table - working closely with CEN and CENELEC groups and the European Commission to shape harmonized standards for the AI Act. Our focus? Building a robust AI Security Standard and solving some of the field's toughest challenges: - How can we confidently protect against data poisoning? - What does future-proof testing for evasion attacks look like? - How do we define countermeasures clearly, so there's no ambiguity for vendors and assessors? - And how do we balance clear guidance with the flexibility needed to support innovation? This is probably the biggest puzzle of my career. A heartfelt thanks to my peers at CEN/CENELEC—I’m constantly learning from you. To the incredible experts at the OWASP AI Exchange, thank you for your invaluable contributions. And to my team at Software Improvement Group, I’m grateful for the opportunity to share our research and insights to make a real difference. Let’s work together to make AI a force for good, and not the ideal target for malicious actors - which it currently is. In the picture, at the left hand, Annegrit Seyerlein-Klug Convenor, At the close right, Wei Wei, Editor, and me on the left, second row, co-editor, at the table with other representatives of WG5 and the European Commission. #ai #aisecurity #aiact

How can this possibly be free? Join some of the best with Rob van der Veer, Mackenzie Jackson, Tiago Teles, Philippe De Ryck, Olle E Johansson, Spandan Chandra, Irfaan Santoe, Robin van Loon, and Jim Manico later this month at OWASP BeNeLux Days Register now: https://lnkd.in/er67-kRf

How can we control AI security with timely, clear, comprehensive, and fair rules? I explored this question during my opening keynote at the annual?ETSI?Security Conference in Sophia Antipolis, France, representing Software Improvement Group. In his welcome message, Director General Jan Ellsberger emphasized the importance of preventing fragmented efforts - a sentiment we all share. Yet, the world hasn't found a way to unify standards effectively for practitioners. OpenCRE is part of the solution, connecting different standards, but the broader landscape remains fragmented. To address this, I’m committed to bringing experts together to tackle the urgent challenge of AI security. If we don’t, AI—due to its ubiquity— becomes the ideal attack vector for malicious actors. Securing AI is a new area for many, which makes our task even more critical. So, a call to action for all standard makers: work closely with your peers, and if that’s a challenge, at least let your experts engage in the OWASP AI Exchange to collaborate on shared research questions. See my slides: https://lnkd.in/e_y7ytvZ In the keynote, I discuss the fragmented standard landscapes, the development of the EU AI Act security standard, and a brief course on AI security, including data poisoning and indirect prompt injection. Event site: https://lnkd.in/edXrRRfb Let’s collaborate and make a difference, fellow experts. #ai #aisecurity #ETSISEC2024

?? I am pleased to announce that I will be speaking at the OWASP Chapter Germany on "Overview of OWASP AI Exchange: A Comprehensive Guide to AI Security." In this session, I will provide an in-depth overview of the OWASP AI Exchange and its mission. I’ll delve into key aspects of AI security, offering insights and practical guidance for navigating the challenges we face today. ?? Event Details: German OWASP Day : Leipzig on November 12-13, 2024! I look forward to an engaging discussion and the opportunity to connect with fellow professionals dedicated to enhancing AI security. https://god.owasp.de/2024/ OWASP AI Exchange OWASP Germany #AIsecurity #OWASP #Cybersecurity #MLSecOps #AI

If you're in the neighborhood, come see the session on security and GenAI by our very own Niklas Bunzel. #ai #genai #aisecurity

I feel incredibly grateful for the opportunity to help design our future with AI, in a whirlwind of a time! ?? Monday: Keynote on AI security for Siemens Industrial Software. ?? Tuesday: Two podcasts with Caitlin Begg on AI adoption. ?? Wednesday: Incredible launch party of ’Luna and the Magic AI Paintbrush’ with co-author, publisher, and superwoman Bessie Schenk. ?? Thursday: Book signing at the World Summit AI and a talk on our future with AI. ?? Friday: Lead a workshop day on AI strategy with a client. And that’s just the start. Next week, I’m looking forward to: ?? Monday: Opening the ETSI security conference in Cannes, advocating broad collaboration on AI security standards. ?? Tuesday: Launching the Software Improvement Group AI Readiness Guide! ?? Wednesday: Article on the partnership I set between CEN and CENELEC - OWASP? Foundation, and how it positions the OWASP AI Exchange. ?? Thursday: Panellist at the European Cyber Security Organisation (ECSO) high-level event in Brussels. ?? Friday: Security event with Brenno de Winter who has another brilliant idea. Thank you to everyone who has listened, participated, and engaged with my work. Your input and feedback make this journey worthwhile. I’m just a guy trying to make sense of it all. As I shared in my summit talk, we all wonder: What is our future with AI? The best answer is, “We don’t know” because it keeps us both curious and cautious. In the end, WE decide: - We decide if we’ll use AI to attend and summarize our meetings. - We decide if we’ll drive in cars that almost do everything right. - We decide how we collaborate with AI while maintaining our skills. -We decide if we’ll put AI in weapons, knowing the risks of data poisoning and evasion attacks. We should explore and make mistakes. But we must also be willing to step back, reflect, and help others do the same. Thank you for joining me on this journey. I should also mention Luc Brandts, Wouter Knigge, Olivia Gambelin, Petri Myllym?ki, Yiannis Kanellopoulos, Peter Boersma, Birgit Geiberger, Paul El Khoury, Arman Iranfar, Huub Janssen, Florence Mottay, Pavel Cholakov, Meindert Kamphuis, and Franjo Bartels. #ai #responsibleai #aisecurity