European Data Compliance Network EUDCN

I feel very honoured to have been nominated twice and finally also shortlisted among 5 finalists for the prestigious (new) Cyber Security Privacy Professional of the Year by the Belgian Cyber Security Coalition, among my esteemed colleagues. Being DPO in a large organisation is very exciting but also challenging, a role that is not possible without the support of top management, Benoît Gilson - CEO Infrabel, Philippe Denayer - CEO TUC RAIL, Naïma Atori - COO TUC RAIL, and their vision and strategy for cyber security and data protection. More info: Belgium’s Cyber Security Awards - Home 2024 Shortlist (cybersecuritycoalition.be) - Award ceremony on December 4, 2024. #dataprotection #privacy #cybersecurity #cybersecuritycoalition

CFOs can help their organizations decarbonize and meet reporting requirements by using consistent metrics and improving accountability. https://mck.co/3Uv1DYN #BeyondWhy

The U.S. Department of Homeland Security and the Cybersecurity and Infrastructure Security Agency are taking steps to facilitate the eventual transition to post-quantum cryptography, as outlined in Secretary of Homeland Security Alejandro N. Mayorkas’ March 2021 vision for #cybersecurity resilience. CISA has also examined post-quantum threats to National Critical Functions (NCFs) and contributed to guidance on the migration to post-quantum #cryptography. This document builds on these efforts by examining the specific risks that #quantum computing could pose to industrial control systems (ICS) and other operational technology (#OT).

🇳🇱 Cybercheck: Beware of supply chain risks! A guide for mapping and controlling supply chain risks for products and services from countries with an offensive cyber programme #tprm #risk #supplychain #cybersecurity

🔥 New technical research publication about Bias in AI 🔥 🚀 𝗔𝗱𝘃𝗮𝗻𝗰𝗶𝗻𝗴 𝘁𝗵𝗲 𝗳𝗶𝗲𝗹𝗱 𝗼𝗳 𝗯𝗶𝗮𝘀 𝗱𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗺𝗶𝘁𝗶𝗴𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗮𝗻𝗱 𝗧𝗿𝗮𝗱𝗶𝘁𝗶𝗼𝗻𝗮𝗹 𝗔𝗜 𝗠𝗼𝗱𝗲𝗹𝘀 On 8th October, we had the pleasure of presenting the results of our technical research at the University of Amsterdam (UvA) together with the six talented student-researchers from the UvA Master in Artificial Intelligence. This journey has been a rewarding yet intense year of collaboration, and today, we are excited to share that our white paper with the results of our research including access to the research materials, code and 3 synthetic datasets is available to everyone! Here’s what you will find in the white paper: 🔹 Three LLM bias studies: we assess whether LLMs reduce or amplify bias in hiring, compare in-processing and post-processing techniques to assess gender bias, and identify novel mechanisms to identify how bias is encoded within LLMs. 🔹 Federated Learning: we explore the balance between privacy and fairness in decentralized systems. 🔹 Causal Fairness in ML: we find novel causality-driven methods to enhance bias identification and mitigation, focusing on causal relationships rather than simple correlations. 🔹 Profile-Based Subgroup Discovery (PSD): we introduce a new approach to uncovering hidden biases within subgroups. These findings showcase both the strengths and limitations of current methods while opening the door for further research and application. 🙌 Explore the white paper and help us continue this crucial work in responsible AI. 𝗪𝗵𝗶𝘁𝗲 𝗽𝗮𝗽𝗲𝗿: https://lnkd.in/eiJwMjTW 𝗙𝘂𝗹𝗹 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵: https://lnkd.in/eZcfZhtp 𝗥𝗵𝗶𝘁𝗲 𝗞𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲𝗯𝗮𝘀𝗲: https://lnkd.in/e-WMEhkc #AI #BiasMitigation #LLMs #FederatedLearning #ResponsibleAI #CausalFairness #AIResearch #Whitepaper #EthicalAI #BiasDetection #Collaboration

Murky Consent: An Approach to the Fictions of Consent in Privacy Law https://lnkd.in/eWvmqyzu

Luxembourg is preparing for DORA 🌐    The recent DORA readiness survey conducted by the Commission de Surveillance du Secteur Financier (CSSF) shows encouraging progress, with 90% of financial entities completing their gap analysis ahead of the January 2025 deadline. While there are challenges ahead, such as fine-tuning resource allocation and governance structures, many entities are well on their way to full compliance.    Currently, 71% of entities perceive themselves as partially ready, while 23% feel mostly ready. This progress is consistent across various entity types, showing strong momentum across the board, with just 6% of entities still in the early stages.    For businesses in the preparatory phase, the coming months offer an opportunity to solidify their digital operational resilience and ensure readiness for DORA’s implementation, which will enhance the security and stability of the financial sector.    See the report here:  https://lnkd.in/eHq6gzEf    #DORA #DigitalResilience #RegulatoryCompliance #FinancialServices #CSSF #Luxembourg

Navigating Generative AI in EU Law: Key Challenges & Recommendations  Generative AI, particularly large language models (LLMs), is transforming industries, yet raises critical legal and ethical questions. 👇👇👇 Here’s a look at the document's insights on how EU law is adapting: 👉 Liability: New proposals like the AI Liability Directive (AILD) and Product Liability Directive (PLD) aim to define who’s accountable when AI causes harm. However, AI’s complexity and unpredictability mean these frameworks may need further refinement. 👉 Privacy: GDPR compliance with AI faces hurdles like data leaks and “hallucinations” (inaccurate outputs). Calls for privacy-preserving techniques and a right to data erasure emphasise the need to better protect users’ data in AI systems. 👉 Intellectual Property: How AI uses data raises copyright concerns. EU law must clarify how AI can access copyrighted materials for training, and whether AI-generated outputs can be protected under IP law. 👉 Cybersecurity: Stronger cybersecurity is vital. The EU’s Cyber Resilience Act and proposed AI Act updates aim to protect AI systems, including from adversarial attacks, misinformation, and potential misuse. The document below highlights that effective AI governance in the EU will require adapting our current laws to meet AI’s unique capabilities and challenges. A more comprehensive framework can empower innovation while ensuring ethical, secure, and compliant AI deployment.