Resaro

𝐁𝐢𝐠 𝐧𝐞𝐰𝐬: Resaro 𝐋𝐚𝐮𝐧𝐜𝐡𝐞𝐬 𝐢𝐭𝐬 𝐄𝐮𝐫𝐨𝐩𝐞𝐚𝐧 𝐎𝐟𝐟𝐢𝐜𝐞 & 𝐖𝐞𝐥𝐜𝐨𝐦𝐞𝐬 𝐋𝐞𝐚𝐝𝐞𝐫𝐬𝐡𝐢𝐩 𝐀𝐝𝐝𝐢𝐭𝐢𝐨𝐧𝐬 We are delighted to announce the opening of Resaro's new co-headquarters in Munich, Germany 🇩🇪, marking a pivotal step in our mission to lead the global AI assurance and testing space. This expansion reinforces our commitment to building trust in AI systems and strengthens our presence in Europe’s vibrant technology ecosystem. We are equally excited to welcome Olly Salzmann as the CEO of Resaro (Germany), and Sebastian Hallensleben as our Chief Trust Officer. Both leaders bring exceptional expertise and vision to Resaro, positioning us for even greater impact across the AI assurance landscape. Working together with April Chin, CEO of Resaro (Singapore), they will build a strong bench of AI testing and evaluation tools and services to enable the AI market to be worthy of trust. Read more in our latest article.

During the last few days we heard a lot about regulations vs technology innovations and the drift in our core values of democracy and free speech between US and Europe at the #MSC25.   Personally, this even more encourages me in my firm conviction, that we have to strengthen the European Union by fostering collaboration within but also with other strategic countries. In order to protect our values and prosperity, we need to accelerate the adoption and procurement of technology innovation into our fundamental capabilities and internal operations as well as create trust and compliance in the usage of disruptive technology.   In our two breakfast roundtables at the Munich Security conference with representatives from the defense and national security ecosystem of Germany, Poland, Europe and Singapore as well as researchers, defense accelerators and dual use AI startups/scale up for defense, we discussed various threat scenarios and how emerging technology in alignment with defense actors and in collaboration with other countries shall respond.   Reflecting on the conversation and listening to expert, I’m very enthused to say that on a talent and operational level, we have everything we need to combat current threats and ongoing attacks as currently happening under water (damaged sea cables in the North Sea), on the ground (Ukraine), in space (drone surveillance at our eastern border) and in the cyber world (bots, deepfake and misinformation). In particular it was very impressive to listen and learn from our guest from #Singapore how they work on those challenges and the pragmatic approach they apply to identify, acquire and integrate emerging dual use technology into their core capabilities for defense and society. Together with my colleauges April Chin, Miguel Fernandes and Sebastian Hallensleben, we will contribute to this mission and thus enhance the collaboration between Singapore and Germany to accelerate the adoption of dual use AI technology into mission critical but core operations and processes. #ai #technology #deterrence #dualuse #singapore #germany #europe #regulation #innovation #trust

📣 𝗦𝗽𝗼𝘁𝗹𝗶𝗴𝗵𝘁 𝗼𝗻 Resaro: 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗧𝗿𝘂𝘀𝘁𝘄𝗼𝗿𝘁𝗵𝘆 𝗔𝗜 𝗮𝗻𝗱 𝗖𝗼𝗻𝘁𝗿𝗶𝗯𝘂𝘁𝗶𝗻𝗴 𝘁𝗼 𝗢𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝘄𝗶𝘁𝗵 𝗔𝗜 𝗩𝗲𝗿𝗶𝗳𝘆! As AI adoption accelerates, ensuring responsible and safe AI development is paramount. We're proud to spotlight Resaro, a premier member of the Foundation for their work in advancing responsible AI and enhancing open-source tools. They've successfully utilised AI Verify to evaluate third-party AI models and offer testing services. Resaro also worked with IMDA and Foundation to streamline the installation, use, and maintenance of test modules: 1. Simplified AI Verify toolkit architecture for modular testing 2. Enhanced support for computer vision frameworks like PyTorch and TensorFlow Read more about Resaro’s experience in contributing to AI Verify toolkit and its insights on AI testing: https://lnkd.in/gDZU4kkd   Join us in contributing to open-source development today!

At the AI Action Summit in France, Mrs Josephine Teo (Minister for Digital Development and Information) announced three initiatives reflecting Singapore’s commitment to rally industry and international partners towards concrete actions that advance #AI #safety. Among these initiatives is the Global AI Assurance Pilot by the AI Verify Foundation — a testbed that matches AI assurance and testing providers with those who need them. Real clients, real use cases, to create testing norms and best practices to drive innovation. 🚢 And we’re glad to come on board. With this Pilot, we’re laying the groundwork for what "good” looks like for #GenAI. Check out this article for more: https://lnkd.in/gsYTvrpE

F. Scott Fitzgerald once said: “The test of a first-rate intelligence is the ability to hold two opposed ideas in mind at the same time and still retain the ability to function." Glad to have met folks at the AI Safety Connect in Paris who believe that innovation and safety can and should co-exist. Olly Salzmann and I got to share about Resaro’s approach to AI application testing in the aviation industry. Also fascinating to see different approaches by PRISM Eval, lucid computing, Conjecture and Infinitio AI to tackle AI safety. We need to continue to bring the different industry, policy, technical and entrepreneurial voices to the table to find actionable solutions for AI safety and innovation. Kudos to Cyrus Hodes, MBR School of Government and Future of Life Institute (FLI) for kicking off this conversation!

The AI Safety and Security Ecosystem is Ready but where is the Market?   The depth and breadth of the Safety community whether from a technical or standards and policy perspective is significant. Whether from frontier model providers, AI research, startups, government-initiated AI Safety Institutes, Offices and Networks.   Now it is time to adapt and create the foundation for an AI safety market. It is inevitable that AI needs safety and security. As Stuart Russell used the analogy in context of working with Artificial General Intelligence “If you want to build a nuclear power plant in the middle of Paris – than the plant needs to be safe and secure before going live”   The supply chain of AI technologies is complex to manage with compounding risk vectors across 3rd party vendors, open-source projects and in-house developments. We strongly believe that the development of rigorous ‘AI assurance’ mechanisms to address unintentional failures as well as malicious misuse across the AI supply chain is becoming a strategic capability of enterprises and governments and thus NOT just a regulatory topic.   Thanks to Cyrus Hodes and team for bringing together so many global experts and hosting the AI Safety Connect side event during the AI Action Summit in Paris. Please find further information about the event and speakers in the comments. Looking forward to continuing and supporting your initiative in the future.   #AISI #aiactionsummit #market #safety #impact #global

We are thrilled to announce the launch of the Singapore Standard SS ISO/IEC 42001, an identical adoption of the international standard ISO/IEC 42001. This groundbreaking standard specifies requirements for establishing, implementing, maintaining, and continually improving an Artificial Intelligence Management System (AIMS) within organizations. It is designed for entities providing or utilizing AI-based products or services, ensuring responsible development and use of AI systems. A special mention to Mr. Andeed Ma, President of the Risk and Insurance Management Association of Singapore (RIMAS), who serves as the Convenor of Singapore Standard ISO/IEC 42001: 2024 AI Management System. RIMAS has taken the lead in driving this standard forward, showcasing its commitment to advancing risk management and AI governance. Alongside Mr. Ma, we acknowledge the invaluable contributions of the Working Group members from various esteemed organizations: - Dr. Sreejith Balakrishnan - Dr. Yao CHENG - Mr. Gerry Chng - Mr. Benjamin Chua - Mr. Gar Yin Hor - Mr. Sengmeng . 邱胜铭  - Mr. Benjamin Mah These experts, representing organizations such as Risk and Insurance Management Association of Singapore (RIMAS), AIQURIS - A TUV SUD Venture Pte Ltd, IMDA, Resaro, SGTech, and ST Engineering, have played a pivotal role in shaping this standard. The relationship between ISO 31000:2018 and ISO 42001 is fundamental, as ISO 31000 serves as a key reference for ISO 42001. This integration underscores the importance of robust risk management principles in AI governance. The framework within ISO 42001 emphasizes systematic risk assessment, dynamic risk management, and comprehensive documentation, ensuring that organizations can navigate the complexities of AI implementation while maintaining appropriate controls and safeguards. Join us for the launch seminar to hear from industry practitioners and experts on how your organization can implement this standard to manage risks and promote the responsible use of AI. Date:19 February 2025 (Wed)   Time: 9.00 am to 12.30 pm   Venue: York Hotel Singapore  The seminar is free-of-charge, but seats are limited. Register early to secure your spot: [https://lnkd.in/g5JMcT7T] Singapore standards (SS) can be purchased in the SS shop: https://lnkd.in/gG8hxScg Let’s embrace this milestone in AI governance together! #AI #ArtificialIntelligence #RiskManagement #ISO42001 #AIMS #Innovation #Governance #SingaporeStandard #RIMAS

Coming up at the #AIActionSummit in Paris 🇫🇷: Resaro’s Chief Trust Officer and Chair of CEN-CENELEC JTC 21 on AI, Sebastian Hallensleben, will take the stage to discuss "𝐓𝐡𝐞 𝐑𝐨𝐥𝐞 𝐨𝐟 𝐂𝐨𝐦𝐦𝐞𝐫𝐜𝐢𝐚𝐥 𝐀𝐈 𝐀𝐬𝐬𝐮𝐫𝐚𝐧𝐜𝐞", while our CEO (Singapore), April Chin, will be part of a “𝐐&𝐀 𝐨𝐧 𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐆𝐥𝐨𝐛𝐚𝐥 𝐀𝐈 𝐒𝐚𝐟𝐞𝐭𝐲 𝐁𝐞𝐧𝐜𝐡𝐦𝐚𝐫𝐤𝐬”. This is part of efforts by the AI Verify Foundation and MLCommons to convene leading practitioners and offer perspectives on the future of #AITesting, #AIAssurance, and #AISafety. 📅 11 February, 1000 - 1200 📍Founders’ Cafe, STATION F Quality and governance, in our view, are inseparable. For an AI system to perform as intended, it must also inspire justified trust in those who use it. This trust comes from rigorous, domain-specific testing with contextual metrics that go beyond baselines, addressing the unique requirements of each use case. Special thanks to AI Verify Foundation and Shameek Kundu for having us. On se voit bientôt, Paris!

  We are invited to participate in the Europol Industry and Research Days 2025.  Europol sits at the forefront of law enforcement innovation and research, and the Agency is responsible for proactive engagement and close collaboration with the private sector, academia, and research institutes.   The level of engagement, professionalism and dedication among the partitioners from Europol, Law Enforcement from EU Member States and EU stakeholders in using technology to face emerging threats was impressive.   We at Resaro could contribute with our solution #clarity - hardening the security infrastructure against deepfakes for our defense units and national security agencies. Clarity offers a flexible but contextual and scalable workflow to apply synthetic generation methods to test and assess various deepfake detector solutions based on the attack domain and type of data and model. Together we address the main pain points in evaluating deepfake detectors.:   👉 There is no one-size-fits-all deepfake detector 👉 Detection technologies are less effective with newer generation algorithms 👉 High benchmark scores do not guarantee real-world performance     Let’s work together to ensure AI is worthy of trust.   #lawenforcement #artificalintelligence #evaluation #trust  

🤖🍪 This week’s treat: a 2-part article series on #deepfakes. The malicious use of deepfakes is raising alarm bells globally—disinformation, fraud, and more. In a world full of "counterfeit people," as philosopher Daniel Bennett calls it, who can we trust online? Regulations, detection tools, and other measures are stepping up to fight back. Read more: https://lnkd.in/ghsSFxja As deepfakes get more sophisticated, detectors should be generalisable across different datasets and generation methods. In our follow-up, we assessed the generalisability of 3 open-source deepfake detectors using the DeepAction Dataset. Dive into it: https://lnkd.in/gstvitvr