We are excited to spotlight the work of Simson Garfinkel and announce his upcoming book, Differential Privacy! Simson researches and writes at the intersection of AI, privacy, and digital forensics. Simson researches and writes at the intersection of AI, privacy, and digital forensics. He is a fellow of the AAAS, the ACM and the IEEE. He earned his PhD in Computer Science at MIT and a Master of Science in Journalism at Columbia University. His new book, Differential Privacy, will be published by MIT Press on March 25, 2025.
Read on for a Q&A with Simson about his new book— and an invitation to celebrate with the OpenDP community!
Q: Tell us about your background in differential privacy.
I was the lead computer science on the development and deployment of differential privacy for the 2020 Census. I came into the Census Bureau unconvinced that differential privacy was the right choice for the Bureau’s next-generation data protection technology. So I immersed myself in the math and the public policy issues, as well as the work that had already been done, and changed my mind: I became a differential privacy convert. Like any new convert, I then tried to share my awakening with others. In doing so, I also became one of the Bureau’s public faces about the need to integrate DP in the data release pipeline of all the Census Bureau’s data products.
I left the Census Bureau in 2021 to work on the Data Inventory Program at the US Department of Homeland Security. The following year I took a temporary break from full-time government service, but I remained active in the DP community.
Q: What motivated you to write this book on differential privacy?
MIT Press approached me in 2023 to see if I would be interested in writing a book about differential privacy for their Essential Knowledge Series. The idea of the series is to offer “concise, beautifully produced books on topics of current interest.” Each of the books are designed to be readable in just a few hours, and MIT Press has developed a template for the books that makes it easy for the reader to come up to speed on a topic. The Press was eager to publish a book on differential privacy because the topic is positioned at the intersection of data science, privacy, and public policy — traditionally all areas that are of concern at MIT.
I would not have been able to write the book without the generous grant that I received from the Alfred P. Sloan foundation. It was important for me that this book be open access, and I was able to use this grant to cover the open access fee. The grant also paid for my illustrator, fact checker, copy editor, and many of the unanticipated expenses that happen when working on a book project.
Q: What makes this book stand out from what’s currently available?
There are quite a few books about differential privacy, but nearly all of them are filled with fancy math proofs, and many also have a big dose of computer code. That’s great if you are a mathematician or computer programmer, but it’s quite disconcerting for people who are not.
In contrast, this book emphasizes the history and philosophy of differential privacy. Whereas other books emphasize the formal definition of differential privacy — they typically start by reciting the mathematical definition from the 2006 paper — this book stresses what’s known as the semantic interpretation of that mathematical equation.
In fact, the equation doesn’t even appear until the second chapter, and when it does, it’s in a cartoon.
Q: What are some challenges of writing about differential privacy for a general audience?
The biggest challenge is having the courage to leave things out. As a computer security and privacy professional, I’ve based my entire career on the premise that technical details matter. But while that may be true when making recommendations to policy makers, it is not true when explaining a technical issue to a general audience. For most people, the details really do not matter. Thus the challenge: how do you leave details out without writing something that is factually wrong because it’s missing necessary qualifiers — qualifiers that people don’t care about and probably wouldn’t understand without a lot of explanation.
Q: This book incorporates a set of humorous illustrations about differential privacy. How do you think the visuals help convey the message of the book?
My original vision for the book was that it was going to be teaching the reader a lot about the math, and so I wanted to have illustrations that explained things like epsilon and delta and the difference between “pure” differential privacy and zero-concentrated differential privacy. But as I started to work on the book, I realized that a book that dove down into those kinds of technical details was not the book that readers needed.
Readers, I realized, needed a book about the policy issues surrounding differential privacy. And in my experience, a very effective way to get to the heart of policy issues is through editorial cartoons.
It just so happens that I’ve known Ted Rall, an award-winning editorial cartoonist, since I was a graduate student in journalism school in the 1980s. I’ve always wanted to work on a project with Ted, so when this opportunity came along, I asked him if he would like to be involved, and he said “yes!”
The cartoons are important in two ways. First, because they are both visual and funny, they help lock the book’s ideas into a person’s long-term memory. Many people recognize this from their own experience, and there’s a lot of scientific evidence about this effect as well. Ted’s cartoons can also go beyond what is possible to what is impossible, or at least improbable, and make arguments about the subject matter that I can’t make myself.
Q: Where can we find the book when it’s available?
You can pre-order the book today from Amazon. It’s also for sale at a lot of other book sellers; you can get the whole list at the MIT Press website.
OpenDP is excited to celebrate this new educational resource for expanding the DP community! Join us (and meet Simson!) on Tuesday, March 25 from 5:00 – 6:00 PM at the 3rd floor West Lounge of the Harvard Science and Engineering Complex for appetizers, refreshments, and a few free copies of the book.
The event is free, but we encourage you to RSVP here for headcount purposes. We look forward to seeing you there!