Data Protection Practices

Today, National Institute of Standards and Technology (NIST) published its finalized Guidelines for Evaluating ‘Differential Privacy’ Guarantees to De-Identify Data (NIST Special Publication 800-226), a very important publication in the field of privacy-preserving machine learning (PPML). See: https://lnkd.in/gkiv-eCQ The Guidelines aim to assist organizations in making the most of differential privacy, a technology that has been increasingly utilized to protect individual privacy while still allowing for valuable insights to be drawn from large datasets. They cover: I. Introduction to Differential Privacy (DP): - De-Identification and Re-Identification: Discusses how DP helps prevent the identification of individuals from aggregated data sets. - Unique Elements of DP: Explains what sets DP apart from other privacy-enhancing technologies. - Differential Privacy in the U.S. Federal Regulatory Landscape: Reviews how DP interacts with existing U.S. data protection laws. II. Core Concepts of Differential Privacy: - Differential Privacy Guarantee: Describes the foundational promise of DP, which is to provide a quantifiable level of privacy by adding statistical noise to data. - Mathematics and Properties of Differential Privacy: Outlines the mathematical underpinnings and key properties that ensure privacy. - Privacy Parameter ε (Epsilon): Explains the role of the privacy parameter in controlling the level of privacy versus data usability. - Variants and Units of Privacy: Discusses different forms of DP and how privacy is measured and applied to data units. III. Implementation and Practical Considerations: - Differentially Private Algorithms: Covers basic mechanisms like noise addition and their common elements used in creating differentially private data queries. - Utility and Accuracy: Discusses the trade-off between maintaining data usefulness and ensuring privacy. - Bias: Addresses potential biases that can arise in differentially private data processing. - Types of Data Queries: Details how different types of data queries (counting, summation, average, min/max) are handled under DP. IV. Advanced Topics and Deployment: - Machine Learning and Synthetic Data: Explores how DP is applied in ML and the generation of synthetic data. - Unstructured Data: Discusses challenges and strategies for applying DP to unstructured data. - Deploying Differential Privacy: Provides guidance on different models of trust and query handling, as well as potential implementation challenges. - Data Security and Access Control: Offers strategies for securing data and controlling access when implementing DP. V. Auditing and Empirical Measures: - Evaluating Differential Privacy: Details how organizations can audit and measure the effectiveness and real-world impact of DP implementations. Authors: Joseph Near David Darais Naomi Lefkovitz Gary Howarth, PhD

On 13 Nov, the Cybersecurity and Infrastructure Security Agency & the Federal Bureau of Investigation (FBI) released a statement (https://lnkd.in/ezrFy_4j) on the US government's investigation into PRC targeting of telco infrastructure: “PRC-affiliated actors have compromised networks at multiple telecommunications companies to enable the theft of customer call records data, the compromise of private communications of a limited number of individuals who are primarily involved in government or political activity, and the copying of certain information that was subject to U.S. law enforcement requests pursuant to court orders. We expect our understanding of these compromises to grow as the investigation continues." With the investigation ongoing, folks should take basic steps now to protect their personal communications. With gratitude to CISA's Senior Technical Advisor Bob Lord (https://lnkd.in/e-WxWiFF) consider the below steps: - Enable FIDO authentication or FIDO https://lnkd.in/ezzyha7t for email & social media accounts - Migrate off SMS MFA for all other logins. Migrate to FIDO/passkeys if you can, otherwise to an authenticator app - Use a password manager for all passwords. Use a strong pass phrase (https://lnkd.in/ebPpTAU5) for the vault password. - Set a telco PIN to reduce chances of a SIM-swap attack - Update the OS and all apps and turn on auto update Additional tips: 1. Encrypt all text and voice communications (some options): - Signal works well on iPhones & Android phones. - iMessage is great if all your contacts are within the Apple ecosystem, though that’s limiting - Collaboration suites like Google Workspace or Teams can work but don’t always encrypt as you might assume. For example, Teams encrypts data point-to-point, meaning it’s decrypted on Microsoft’s servers before re-encrypting it to the recipient. If you want end-to-end encryption, there’s an option, but it’s off by default and only supports two people on the call. - WhatsApp might be ok for some people based on their threat model but understand metadata it keeps (https://lnkd.in/eQkP-Ety) & how it's used (https://lnkd.in/eiZmxgi4). 2. If you use an iPhone disable these carrier-provided services that increase the attack surface: - Disable: Settings > Apps > Messages > Send as Text Message - Disable: Settings > Apps > Messages > RCS Messaging > RCS Messaging 3. Protect DNS lookups (some options): - Apple iCloud Private Relay - Cloudflare’s 1.1.1.1 resolver - Quad9’s 9.9.9.9 resolver 4. Use recent hardware: Apple (13 or newer) or Google (Pixel 6 or newer) 5. Depending on your threat model, consider enabling Lockdown Mode on iPhones: It will disable some features, but it’s manageable

🚨 AI Privacy Risks & Mitigations Large Language Models (LLMs), by Isabel Barberá, is the 107-page report about AI & Privacy you were waiting for! [Bookmark & share below]. Topics covered: - Background "This section introduces Large Language Models, how they work, and their common applications. It also discusses performance evaluation measures, helping readers understand the foundational aspects of LLM systems." - Data Flow and Associated Privacy Risks in LLM Systems "Here, we explore how privacy risks emerge across different LLM service models, emphasizing the importance of understanding data flows throughout the AI lifecycle. This section also identifies risks and mitigations and examines roles and responsibilities under the AI Act and the GDPR." - Data Protection and Privacy Risk Assessment: Risk Identification "This section outlines criteria for identifying risks and provides examples of privacy risks specific to LLM systems. Developers and users can use this section as a starting point for identifying risks in their own systems." - Data Protection and Privacy Risk Assessment: Risk Estimation & Evaluation "Guidance on how to analyse, classify and assess privacy risks is provided here, with criteria for evaluating both the probability and severity of risks. This section explains how to derive a final risk evaluation to prioritize mitigation efforts effectively." - Data Protection and Privacy Risk Control "This section details risk treatment strategies, offering practical mitigation measures for common privacy risks in LLM systems. It also discusses residual risk acceptance and the iterative nature of risk management in AI systems." - Residual Risk Evaluation "Evaluating residual risks after mitigation is essential to ensure risks fall within acceptable thresholds and do not require further action. This section outlines how residual risks are evaluated to determine whether additional mitigation is needed or if the model or LLM system is ready for deployment." - Review & Monitor "This section covers the importance of reviewing risk management activities and maintaining a risk register. It also highlights the importance of continuous monitoring to detect emerging risks, assess real-world impact, and refine mitigation strategies." - Examples of LLM Systems’ Risk Assessments "Three detailed use cases are provided to demonstrate the application of the risk management framework in real-world scenarios. These examples illustrate how risks can be identified, assessed, and mitigated across various contexts." - Reference to Tools, Methodologies, Benchmarks, and Guidance "The final section compiles tools, evaluation metrics, benchmarks, methodologies, and standards to support developers and users in managing risks and evaluating the performance of LLM systems." 👉 Download it below. 👉 NEVER MISS my AI governance updates: join my newsletter's 58,500+ subscribers (below). #AI #AIGovernance #Privacy #DataProtection #AIRegulation #EDPB

ROPA: MORE THAN COMPLIANCE, A GATEWAY TO REAL DATA PROTECTION The Record of Processing Activities or RoPA is more than just a legal obligation. It is a gateway to building real trust and accountability in how an organization manages personal data. With a well-prepared RoPA a company does not simply comply with Regulation but lays a strong foundation for responsible data protection. Every business unit needs to have the capability to create and maintain a RoPA. It should never be treated as something that only concerns the legal team or the Data Protection Officer. Marketing teams HR IT and operations all handle personal data in some way and each must know how to record check and improve how that data is processed. A good RoPA can serve many purposes at once. It helps teams analyze whether the purpose of processing truly matches the data they collect. It works as a single source of truth for what personal data the organization holds where it comes from who processes it who has access to it and how long it is kept. It is also a clear map of data flows which helps spot risks and strengthen controls. There is no strict format for a RoPA and no specific platform that must be used. What matters most is that it remains practical and flexible. A RoPA should be easy to update whenever something changes whether by adding new information adjusting details or expanding columns and fields as needed. It must also be easy to extract and share in a format that other teams and stakeholders can read and understand. According to Indonesia’s draft PDP Regulation a RoPA should at minimum include the names and contact details of controllers joint controllers and processors the contact for the Data Protection Officer the source and purpose of collection the legal basis the types of personal data and categories of data subjects other parties who can access the data how data subjects’ rights are fulfilled data flow mapping retention periods and technical and organizational measures for protecting personal data. When built and used properly a RoPA is not just a static document for audits but a living record that guides daily actions and builds a culture where protecting personal data is about doing what is right openly and responsibly. Here is an example of a RoPA you can adapt, just add or adjust the questions to match legal requirements or your company’s needs. https://lnkd.in/gKSj9Dzm

RoPA is the paperwork you hate. I get it It’s a one & done, move on with your life situation → or so you think But let me show you something You can lean on your RoPA & it saves you Not during compliance checks - but in real-life situations I’ve done it time & time again so I’ll show you how: 1- Breach You can use it to swiftly locate the source, affected data, systems you need to shut down & your first responder action plan 2- Auditors Auditors love asking for RoPAs. You get fewer questions, it makes them linger less. Everyone’s happy & we can all go back to work 3- DSARs One of our clients, following a breach, was hit with multiple DSARs at once. The only way we were able to avoid a logistical nightmare was going back to that RoPA 4- Retention We once helped a client reduce data storage costs just by getting their RoPA to flag data that needs to be deleted 5- M&A Our client recently acquired another business. Their RoPA was crucial to understanding their data landscape 6- Vendors We’re navigating this right now - our client had a vendor suddenly change terms. We’re looking into their RoPA to identify what data they handled & whether it’s worth renegotiating or just replacing them 7- Privacy notices We’ve helped businesses refresh their privacy notices using their RoPA. They went from generic, vague statements to actually reflecting their practices. If you think a RoPA is optional, think again Knowing your ̉business is power But knowing where your data is, how it flows & who’s using it is priceless A RoPA isn’t a burden → it’s peace of mind P.S. What are some more ways a RoPA can be helpful?

Microsoft AI Teams will soon tell your boss where you are. Starting December 2025, Teams can automatically detect when you connect to your company’s Wi-Fi and update your location to “in the office.” It sounds like a small feature. It isn’t. Location tracking through workplace networks is the newest frontier in digital surveillance, and it’s coming through your collaboration software. Microsoft says the feature is opt-in. That is very good. But, that decision will rest largely with employers and admins, not the average employee trying to meet deadlines. If you work for a Microsoft-using organization, now is the time to ask: Is our company planning to activate this feature? Has consent been properly documented? If you represent a union, this deserves to be on your next agenda. The GDPR and UK Data Protection Act require transparency, necessity, and proportionality for any location tracking. Under the EU AI Act, this may also fall under high-risk processing of biometric and personal data for workplace management. Employers must conduct a fundamental rights impact assessment before rolling it out. This isn’t paranoia. It is risk management, employee rights, and compliance. Workplace tracking without explicit, informed consent can violate privacy law in multiple jurisdictions, and it may open employers to liability under both GDPR and the EU AI Act’s risk provisions. If your organization uses Microsoft Teams with minors, such as schools or training programs, the stakes are even higher. Here’s what to do as an employee, parent, or guardian: 🔹 Ask your IT administrator if “location autodetection” is enabled. 🔹 Request a copy of the company’s Data Protection Impact Assessment (DPIA). 🔹 Ensure opt-in consent is voluntary and revocable. 🔹 Check that logs are deleted regularly and not used for performance evaluation. Transparency is not optional. #DigitalSovereignty #WorkplacePrivacy #AICompliance #GDPR #MicrosoftTeams Image source: SlashGear, https://lnkd.in/di5WvY2e From Microsoft: Microsoft 365 Roadmap: https://lnkd.in/dYc3N9TX Microsoft Learn (Configure auto-detect of work location): https://lnkd.in/dtEkYNqB

I recently talked with a fellow marketer about account scoring in #ABM.  They struggled to adapt to the recent privacy law, which made me think: We must rethink how we do this. Account scoring is evolving. It's moving way beyond simple intent data.  You need a new approach to find these high-potential accounts in a privacy-first world. I call it Account Scoring 2.0. But why? ➖Traditional intent data is becoming less reliable.   ➖Privacy regulations are changing.   ➖Third-party cookies are fading away.   ➖We can't rely on old methods.   ➖We need to be more innovative. 👉🏾 Using Account Scoring 2.0 helps you focus on first-party data. This is data you collect directly from your target accounts like: Website visits. Content downloads.    Engagement within your emails.  Collecting and analyzing this data is valuable. It's also privacy-compliant. 👉🏾 Other things to look out for are behavioral signals.  Look out for target accounts engaging with your content.    Are they attending your webinars?  Are they interacting with your sales team?  These actions show interest and suggest potential. 👉🏾Predictive modeling plays a key role too. You can use AI to analyze first-party and behavioral data.  This helps you predict which accounts are most likely to convert.  It allows you to prioritize your efforts. Remember, it's about working smarter, not harder. 👉🏾 Don't forget contextual data; it matters, too.  What's happening in the market?  Look for industry trends that align with your offerings.  Are there changes in your target accounts' businesses?  Understanding the context helps refine your scoring. Look at Account Scoring 2.0 as a strategy, not just technology. It's more about understanding your ideal customer profile. It's about aligning sales and marketing and building relationships while respecting privacy more efficiently. What are your thoughts on the future of account scoring? Have you used it before? #b2bmarketing #marketingstrategy

Privacy-enhancing technologies like homomorphic encryption, differential privacy, and federated learning are redefining how businesses manage data, proving that safeguarding individual privacy doesn't have to come at the cost of losing meaningful insights. Privacy-enhancing technologies (PETs) are advanced tools that allow secure data processing while safeguarding personal identities. Homomorphic encryption enables computations on encrypted data without decryption, maintaining strict confidentiality. Differential privacy ensures dataset utility by adding controlled noise, preventing the exposure of individual data points. Federated learning decentralizes analysis by keeping sensitive data on local devices, reducing the risks of breaches. These methods balance privacy and usability, ensuring compliance with regulations like GDPR while empowering businesses to leverage data responsibly and ethically. #PETs #Privacy #DataSecurity #EthicalAI #DifferentialPrivacy #HomomorphicEncryption #FederatedLearning #DataProtection

How To Handle Sensitive Information in your next AI Project It's crucial to handle sensitive user information with care. Whether it's personal data, financial details, or health information, understanding how to protect and manage it is essential to maintain trust and comply with privacy regulations. Here are 5 best practices to follow: 1. Identify and Classify Sensitive Data Start by identifying the types of sensitive data your application handles, such as personally identifiable information (PII), sensitive personal information (SPI), and confidential data. Understand the specific legal requirements and privacy regulations that apply, such as GDPR or the California Consumer Privacy Act. 2. Minimize Data Exposure Only share the necessary information with AI endpoints. For PII, such as names, addresses, or social security numbers, consider redacting this information before making API calls, especially if the data could be linked to sensitive applications, like healthcare or financial services. 3. Avoid Sharing Highly Sensitive Information Never pass sensitive personal information, such as credit card numbers, passwords, or bank account details, through AI endpoints. Instead, use secure, dedicated channels for handling and processing such data to avoid unintended exposure or misuse. 4. Implement Data Anonymization When dealing with confidential information, like health conditions or legal matters, ensure that the data cannot be traced back to an individual. Anonymize the data before using it with AI services to maintain user privacy and comply with legal standards. 5. Regularly Review and Update Privacy Practices Data privacy is a dynamic field with evolving laws and best practices. To ensure continued compliance and protection of user data, regularly review your data handling processes, stay updated on relevant regulations, and adjust your practices as needed. Remember, safeguarding sensitive information is not just about compliance — it's about earning and keeping the trust of your users.

Isabel Barberá: "This document provides practical guidance and tools for developers and users of Large Language Model (LLM) based systems to manage privacy risks associated with these technologies. The risk management methodology outlined in this document is designed to help developers and users systematically identify, assess, and mitigate privacy and data protection risks, supporting the responsible development and deployment of LLM systems. This guidance also supports the requirements of the GDPR Article 25 Data protection by design and by default and Article 32 Security of processing by offering technical and organizational measures to help ensure an appropriate level of security and data protection. However, the guidance is not intended to replace a Data Protection Impact Assessment (DPIA) as required under Article 35 of the GDPR. Instead, it complements the DPIA process by addressing privacy risks specific to LLM systems, thereby enhancing the robustness of such assessments. Guidance for Readers > For Developers: Use this guidance to integrate privacy risk management into the development lifecycle and deployment of your LLM based systems, from understanding data flows to how to implement risk identification and mitigation measures. > For Users: Refer to this document to evaluate the privacy risks associated with LLM systems you plan to deploy and use, helping you adopt responsible practices and protect individuals’ privacy. " >For Decision-makers: The structured methodology and use case examples will help you assess the compliance of LLM systems and make informed risk-based decision" European Data Protection Board