Data Science and Privacy - sugarcoated or straight up? It Depends (with Katharine Jarmul of Cape...

Send us Fan Mail (https://www.buzzsprout.com/840448/fan...) Privacy and data protection are not just a job for lawyers or professionals who specialize in privacy - not anymore. Technology plays an important role in ensuring personal data can remain private. Ensuring that personal data is secure but useful requires a level of skill found in data scientists. In this episode of Serious Privacy (https://seriousprivacy.buzzsprout.com/) , Paul Breitbarth (  / paulbreitbarth  ) and K Royal (  / kroyal  ) searched for just such a skilled individual,Katharine Jarmul (  / katharinejarmul  ) , the Head of Product at Cape Privacy (https://capeprivacy.com/) , and a data scientist. Cape Privacy is a New York-based company assisting others with machine learning, data security and adding value to data. Katharine explains what data science actually is, how to keep data private, useful and valuable at the same time, and how to create synthetic data (https://datasciencecampus.ons.gov.uk/...) appropriately. Also a big question when it comes to powerful technology revolves around the ethics (https://plato.stanford.edu/entries/et...) and the investment of individual technologists in the ethics of privacy. Join us as we discuss these topics and more, such as GPT-3 (https://en.wikipedia.org/wiki/GPT-3) , “this person does not exist (https://thispersondoesnotexist.com/) ,” the work of Cynthia Dwork (https://datascience.harvard.edu/peopl...) , and differential privacy (https://privacytools.seas.harvard.edu...) vs the generative model (https://openai.com/blog/generative-mo...) . As often happens in an episode, certain topics in privacy are revisited, mainly because they are wicked problems with no identified solution. One such topic Katharine discussed is bias in machine learning and approaches to solving bias once identified. Throughout this episode, we reference quite a few resources that we will provide the links - as always.   Resources • IAPP article on AI and synthetic data: https://iapp.org/news/a/accelerating-... • Federated / Collaborative Learning Introduction: https://federated.withgoogle.com/ • Encrypted Learning with TF-Encrypted (also can be used in a collaborative setting where we are sharing data):   / encrypted-deep-learning-training-and-predi...   • Europe - Ethics guidelines for trustworthy AI https://ec.europa.eu/futurium/en/ai-a... Social Media Twitter: @privacypodcast, @EuroPaulB, @heartofprivacy, @trustarc, @kjam, @capeprivacy Instagram @seriousprivacy If you have comments or questions, find us on LinkedIn (  / seriousprivacy  ) and Instagram @seriousprivacy, and on BlueSky under @seriousprivacy.eu, @europaulb.seriousprivacy.eu, @heartofprivacy.bsky.app and @igrobrien.seriousprivacy.eu, and email [email protected] (mailto:[email protected]) . Rate and Review us! From Season 6, our episodes are edited by Fey O'Brien. Our intro and exit music is Channel Intro 24 by Sascha Ende (https://filmmusic.io/) , licensed under CC BY 4.0. with the voiceover by Tim Foley.