Privacy and Cybersecurity Roundup Covers EU Rules, AI Risks and CISA Reporting

Summary
The roundup covers EU consumer and data-protection developments, guidance on secure data environments and public-sector LLMs, NIST’s taxonomy of AI attacks and mitigations, and CISA’s enhanced voluntary incident-reporting form.
Key points
- The European Commission proposed a Digital Fairness Act covering dark patterns, profiling-based advertising and pricing, addictive design, subscriptions, and AI-based purchasing agents.
- Final EDPB guidance says legitimate-interest processing, including for fraud prevention and information security, requires a documented necessity and balancing assessment; security monitoring must be proportionate.
- Spanish and Greek data protection authorities recommend controls for secure processing environments, including access restrictions, logging, audits, oversight of cloud providers, and measures to prevent data export.
- Hamburg’s LLMoin case study describes safeguards for a public-sector LLM deployment, including input and output controls, RAG, red teaming, and human review; the system does not make legal decisions.
- NIST’s draft adversarial machine learning taxonomy addresses agentic AI risks such as agent hijacking, memory poisoning, malicious tools, and resource exhaustion, while warning that mitigations have limitations.
- CISA’s updated incident-reporting form adds more detailed incident questions, including MITRE ATT&CK terminology; it does not create new mandatory reporting obligations.
Article Details
- Event Type
- Cybersecurity and data-protection policy, guidance, and reporting updates
- Impact
- The European Commission proposed a Digital Fairness Act that would add consumer protections for online interfaces, profiling-based advertising and pricing, influencer marketing, digital engagement features, and subscriptions. Data-protection authorities issued guidance on legitimate interests, secure processing environments, location-based communications, and GDPR-compliant LLM use. NIST published a draft taxonomy of attacks against machine-learning and agentic AI systems. CISA updated its incident-reporting form to collect more structured information; the update does not create new mandatory reporting obligations.
Vendors
FastrMobi AB(IMY-2026-2532), presents the conclusions of IMY’s regulatory innovation sandbox project with FastrMobi AB, a company developing technology that allows businesses to communicate with individuals based on theirMicrosoftLLMoin uses an OpenAI model accessed through a Microsoft Azure environment under a data processing arrangement. Rather than developing or retraining its own model, Hamburg has implemented an application layer thatOpenAILLMoin uses an OpenAI model accessed through a Microsoft Azure environment under a data processing arrangement. Rather than developing or retraining its own model, Hamburg has implemented an application layer that
Products
LLMoin(HmbBfDI) published a case study examining the development, regulatory assessment, and deployment of LLMoin, an AI chatbot designed for use in Germany’s public administration. The document explains howMicrosoft AzureLLMoin uses an OpenAI model accessed through a Microsoft Azure environment under a data processing arrangement. Rather than developing or retraining its own model, Hamburg has implemented an application layer that
Countries
Germanythe development, regulatory assessment, and deployment of LLMoin, an AI chatbot designed for use in Germany’s public administration. The document explains how organizations can implement large language modelsGreece🇪🇸🇬🇷 Spanish and Greek DPAs Publish Joint Guidance on Secure Processing EnvironmentsSpain🇪🇸🇬🇷 Spanish and Greek DPAs Publish Joint Guidance on Secure Processing EnvironmentsSweden🇸🇪 Sweden’s IMY Issues Guidance on Privacy-Compliant Location-Based Mobile CommunicationsUnited StatesOn November 9, 2026, the U.S. National Institute of Standards and Technology (NIST) released an initial public draft of NIST SP 2400-102e2026, Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and
Industries
digital advertisingprovisions would affect online retailers, application developers, subscription-based businesses, digital advertising providers, gaming companies and other traders serving EU consumers, including businessesGamingonline retailers, application developers, subscription-based businesses, digital advertising providers, gaming companies and other traders serving EU consumers, including businesses established outside the EU.Public administrationregulatory assessment, and deployment of LLMoin, an AI chatbot designed for use in Germany’s public administration. The document explains how organizations can implement large language models (LLMs) in accordance
