Who should I follow to learn about Privacy-Preserving AI from leading experts?

If you want to learn **Privacy-Preserving AI** (PPAI), I’d follow a mix of researchers, practitioners, and organizations across **differential privacy, federated learning, secure computation, confidential computing, and privacy in LLMs**. Here are strong people to start with: ## Core researchers to follow - Cynthia Dwork - One of the founders of **differential privacy**, the mathematical foundation behind many modern privacy guarantees in AI. Start with her work if you want the theory behind privacy budgets, noise mechanisms, and formal guarantees. - Nicolas Papernot - A leading researcher at the intersection of **privacy, security, and machine learning**, including differential privacy, privacy attacks, and trustworthy ML. [Nicolas Papernot - HomePage](https://www.papernot.fr/?utm_source=chatgpt.com) - Yves-Alexandre de Montjoye - Known for work on privacy risks in large-scale data, re-identification, and privacy-preserving technologies. - Vitaly Shmatikov - Follow for privacy leakage, ML security, and attacks against AI systems. - Peter Kairouz - A key figure in federated learning and privacy-preserving ML research. - Virginia Smith - Strong for understanding federated learning systems and their practical limitations. [Qi Pang](https://www.cs.cmu.edu/~qpang/?utm_source=chatgpt.com) - Li Xiong - Works on privacy-preserving data sharing, federated learning, and healthcare AI applications. [Li Xiong (computer scientist)](https://en.wikipedia.org/wiki/Li_Xiong_%28computer_scientist%29?utm_source=chatgpt.com) - T. Tony Cai - Good to follow for the statistical foundations of differential privacy and federated learning. [T. Tony Cai](https://tony-cai.com/?utm_source=chatgpt.com) ## Privacy + LLMs (very relevant now) - Reza Shokri - Known for membership inference attacks and understanding how models leak training data. - Huseyin Inan - Works on privacy-preserving language models and LLM-related privacy techniques. [Privacy Preserving Machine Learning Innovation - Microsoft Research: Publications](https://www.microsoft.com/en-us/research/group/privacy-preserving-machine-learning-innovation/publications/?utm_source=chatgpt.com) - Florian Tramèr - Excellent for understanding model extraction, privacy attacks, and AI security. ## Industry teams worth following - [Microsoft Research Privacy-Preserving Machine Learning Innovation](https://www.microsoft.com/en-us/research/group/privacy-preserving-machine-learning-innovation/publications/?utm_source=chatgpt.com) - Strong output on differential privacy, private synthetic data, private LLM training, and federated learning. [Privacy Preserving Machine Learning Innovation - Microsoft Research: Publications](https://www.microsoft.com/en-us/research/group/privacy-preserving-machine-learning-innovation/publications/?utm_source=chatgpt.com) - [Google Research](https://research.google/?utm_source=chatgpt.com) - Major contributor to differential privacy, federated learning, and privacy-enhancing technologies. - [OpenMined](https://www.openmined.org/?utm_source=chatgpt.com) - Very practical if you want tutorials, open-source tools, and implementation-focused learning. - [TensorFlow Privacy](https://www.tensorflow.org/responsible_ai/privacy?utm_source=chatgpt.com) - Useful for learning how differential privacy is applied in real ML pipelines. ## Communities and conferences Follow these: - Privacy Enhancing Technologies Symposium (PETS) - One of the best places for privacy research. - International Conference on Learning Representations (ICLR) - Increasingly important for private ML and trustworthy AI. - NeurIPS - Many privacy-preserving ML papers appear here. - ACM Conference on Computer and Communications Security (ACM CCS) - Great for privacy attacks and defenses. ## A learning path I’d recommend 1. **Differential privacy basics** - Cynthia Dwork’s papers and talks - [OpenMined](https://www.openmined.org/?utm_source=chatgpt.com) tutorials 2. **Federated learning** - Follow Peter Kairouz, Virginia Smith, and Google Research work 3. **Privacy attacks** - Learn membership inference, model inversion, and data extraction attacks (Papernot, Shmatikov, Tramèr) 4. **Private LLMs** - Look into: - differential privacy for fine-tuning - private synthetic data - secure inference - confidential computing If you want a **short “top 10 people to follow on X/LinkedIn/GitHub” list** or a **beginner-to-expert reading roadmap with papers and courses**, I can make that too.

Here’s a curated list of **leading voices and organizations** you should follow if you want to stay on top of **Privacy-Preserving AI** research, policy, and practice: ### 🌍 Academic & Policy Experts * **Alessandro Mantelero** – Professor of Law & Technology, Polytechnic University of Turin. Focuses on AI, data protection, and digital societies. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) * **Aurélien Bellet** – Research Director at INRIA, specializing in privacy-preserving machine learning. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) * **Amber Sinha** – Senior Fellow at Mozilla Foundation, working on trustworthy AI and privacy. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) * **Byoungpil Kim** – Professor at KAIST, researching privacy in AI systems. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) ### 🏢 Industry Leaders * **Alexandra Ebert** – Chief Trust Officer at MOSTLY AI, focusing on synthetic data and privacy-preserving solutions. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) * **Brian Quirk** – Microsoft expert in AI and privacy policy. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) * **Xabi Uribe-Etxebarria** – Founder & CEO of Sherpa.ai, a company pioneering privacy-preserving AI platforms. [Meet the Team leading Privacy Preserving AI | Sherpa.ai](https://sherpa.ai/about/) * **Tom Gruber** – Chief AI Strategy Officer at Sherpa.ai, co-founder of Siri, working on privacy-first AI innovation. [Meet the Team leading Privacy Preserving AI | Sherpa.ai](https://sherpa.ai/about/) ### 🔬 Research Community * **Kunal Talwar** – Apple ML researcher, known for work on differential privacy and federated learning. [Apple Workshop on Privacy-Preserving Machine Learning & AI 2026 - Apple Machine Learning Research](https://machinelearning.apple.com/updates/ppml-2026) * **Franziska Boenisch** – CISPA researcher, studying privacy risks in foundation models. [Apple Workshop on Privacy-Preserving Machine Learning & AI 2026 - Apple Machine Learning Research](https://machinelearning.apple.com/updates/ppml-2026) * **Salil Vadhan** – Harvard professor, leading work on differential privacy and secure computation. [Apple Workshop on Privacy-Preserving Machine Learning & AI 2026 - Apple Machine Learning Research](https://machinelearning.apple.com/updates/ppml-2026) ### 📚 Organizations & Forums * **OECD.AI Expert Group on AI, Data, and Privacy** – Brings together global policymakers, academics, and industry leaders to shape privacy-preserving AI governance. [Data privacy – Experts - OECD.AI](https://oecd.ai/en/site/data-privacy/experts) * **Future of Privacy Forum (FPF)** – Hosts workshops on privacy-enhancing technologies and AI infrastructure. [FPF Hosts Frontiers Workshop on Privacy, AI, and Emerging Infrastructure](https://fpf.org/blog/fpf-hosts-frontiers-workshop-on-privacy-ai-and-emerging-infrastructure/) * **Sherpa.ai** – A company dedicated to building privacy-preserving AI platforms, recognized with multiple international awards. [Meet the Team leading Privacy Preserving AI | Sherpa.ai](https://sherpa.ai/about/) * **Apple Machine Learning Research** – Runs workshops on privacy-preserving ML, publishing cutting-edge research. [Apple Workshop on Privacy-Preserving Machine Learning & AI 2026 - Apple Machine Learning Research](https://machinelearning.apple.com/updates/ppml-2026) ✅ **Tip for following them:** * On **Twitter/X** , many of these researchers (e.g., Salil Vadhan, Franziska Boenisch) share updates on privacy-preserving AI. * On **LinkedIn** , organizations like Sherpa.ai and FPF regularly post about new projects and workshops. * Academic talks and papers are often shared via **Google Scholar** or **arXiv**. Would you like me to create a **short “who to follow” list on Twitter/X and LinkedIn** with direct handles and pages, so you can easily connect with them?

To master **Privacy-Preserving AI** and Privacy-Enhancing Technologies (PETs) like federated learning, differential privacy, homomorphic encryption, and secure multi-party computation, tracking the right researchers, open-source communities, and organizations is essential. **Key Researchers & Academics** * **Nicolas Papernot:** Professor at the University of Toronto and researcher at Vector Institute. A pioneer in machine learning security, differential privacy, and adversarial robustness. * **Reza Shokri:** Professor at National University of Singapore (NUS) known for foundational work on privacy risks in machine learning, membership inference attacks, and data leakage. * **Carmit Hazay:** Bar-Ilan University researcher specializing in cryptography, secure multi-party computation (MPC), and privacy-preserving protocols. * **Flavio du Pin Calmon:** MIT professor focusing on information theory, data privacy, and algorithmic fairness in machine learning. **Pioneering Organizations & Communities** * **OpenMined:** A massive open-source community dedicated to making AI privacy-preserving through tools like PySyft and PyGrid. Following their community channels and founder **Andrew Trask** offers practical insights into federated learning and encrypted computing. [Privacy-Preserving AI Summary: MIT Deep Learning Series](https://openmined.org/blog/privacy-preserving-ai-a-birds-eye-view/#:~:text=And%20this%20is%20exactly%20what%20the%20OpenMined,data%20we%20can't%20actually%20or%20directly%20see.) * **Stanford Institute for Human-Centered AI (HAI):** Features regular research and commentary on the intersection of AI governance, data privacy policies, and regulatory compliance. * **The Alan Turing Institute (Data Privacy Theme):** A leading UK institute publishing extensive guides and research on ethical and privacy-preserving data science. **Industry Practitioners & Organizations** * **Apple Machine Learning Research:** Apple has long been a pioneer in deploying differential privacy at scale across consumer devices; their published papers and research teams are top-tier sources. * **Google Privacy & Security Research:** The team behind TensorFlow Privacy and foundational federated learning architectures used in products like Gboard. [AI Data Security & Privacy: Protecting Training & Inference Data](https://medium.com/@dave-patten/ai-data-security-privacy-protecting-training-inference-data-8c1267cc446c#:~:text=*%20Open%2Dsource%20tools%20like%20PySyft%20and%20Google's,improve%20autocomplete%20suggestions%20while%20maintaining%20user%20privacy.) What specific pillar of privacy-preserving AI - such as federated learning, differential privacy, or cryptographic methods - are you looking to dive into first?

To truly understand **Privacy-Preserving AI** (encompassing techniques like differential privacy, federated learning, and secure multi-party computation), you should follow leading academics, open-source communities, and industry researchers who actively publish and build in this space. * * * ** _Nicolas Papernot_** - A Professor at the University of Toronto and the Vector Institute, Papernot is a prominent voice in machine learning security and privacy. He leads the CleverHans lab and focuses heavily on how to securely and privately deploy ML in critical fields like healthcare. * * * **_Andrew Trask_** - Known for his pioneering work with OpenMined, Trask is a central figure in democratizing privacy-preserving AI tools (like PySyft). Following him and the OpenMined Blog gives you an incredible window into practical remote execution, federated learning, and differential privacy. * * * **_Cynthia Dwork_** - A professor at Harvard and a foundational pioneer of **Differential Privacy**. While she is more rooted in theoretical computer science, understanding her foundational mathematical framework is mandatory for anyone serious about the field. * * * **_Reza Shokri_** - A professor at National University of Singapore (NUS) whose research focuses extensively on privacy and security in machine learning, membership inference attacks, and data protection in deep learning models. * * * **_Florian Tramèr_** - Assistant Professor at ETH Zurich. He does brilliant work at the intersection of security, privacy, and machine learning, frequently breaking down vulnerabilities in ML models and how to build better defenses. * * * **_Dawn Song_** - A professor at UC Berkeley and founder of Oasis Labs. She is a powerhouse in computer security, blockchain, and privacy-preserving machine learning, focusing heavily on decentralized and confidential computing. [Privacy-Preserving AI Summary: MIT Deep Learning Series](https://openmined.org/blog/privacy-preserving-ai-a-birds-eye-view/)[Andrew Trask | Privacy-Preserving AI & The Future of Data](https://www.youtube.com/watch?v=_ZrLESgZf1Q&t=442) If you'd like, let me know: * Are you looking for **academic papers and theory** or **hands-on coding tutorials**? * What specific privacy technique interests you most (e.g., **Federated Learning** , **Differential Privacy** , or **Homomorphic Encryption**)? I can tailor a specific **learning roadmap** or suggest further resources!