“We will know more about how to be successful a year from now, and those things we learn will probably last and stay with us to be used judiciously in whatever the future of education is.”
- John Mitchell


Professor and Chair of the Computer Science Department at Stanford University

Monica Campbell of PRI's The World Interviews John Mitchell





WHY HIM, WHY NOW? As the former vice provost for online learning at Stanford and a web security expert, he can speak to what the future of online learning will look like post-pandemic, how to balance privacy concerns with access to personal data, and how to improve access and level the playing field among institutions.

EXPERTISE Online learning, programming language, computer security, access control, privacy, web security


Mitchell was the Vice Provost for Teaching and Learning at Stanford from 2015-2018. Before that, he was the Vice Provost for Online Learning from 2012 to 2015 (the position was created for him). In 2009, he built Stanford CourseWare with six undergraduate students, which was the foundation for Stanford’s first flipped-classroom experiments and helped inspire Stanford’s first massive open online courses in 2011. His research focuses on programming analysis and design, improving network security, authorization and access control, web security, and privacy. His work has produced concepts that are used in the Java programming language, improved the security of wireless networking protocols, and helped shape the security architecture of the Google Chrome browser.

DEGREES Bachelor’s (B.S.) from Stanford University Master’s from MIT Ph.D. from MIT


Teaching Online in 2020: Experiments, Empathy, Discovery “How Good Are Humans at Solving CAPTCHAs? A Large Scale Evaluation,” Proc. IEEE Symposium on Security and Privacy, May 2010 CAPTCHAs are designed to be easy for humans to solve but hard for machines to decode (CAPTCHA stands for completely automated public Turing test to tell computers and humans apart — it’s a test used to determine whether or not a user is human). But Mitchell and others found that CAPTCHAs — especially audio ones — are often difficult for humans to solve as well. They found that non-native English speakers especially struggle to complete audio CAPTCHAs and recommended that they thus be overhauled. “A Learning-Based Approach to Reactive Security,” Proc. of the 14th International Conference on Financial Cryptography and Data Security, 2010. Mitchell and others reconsider the conventional opinion that proactive security is superior to reactive security. They demonstrate that reactive security can be as effective as proactive security provided the reactive defender takes a holistic approach to learning from past attacks rather than overreacting to the most recent attack. Books: Theoretical Aspects of Object-Oriented Programming (1994), Foundations for Programming Languages (1996), Concepts in Programming Language (2003)


“What's Next in Online Learning? Stanford Names Vice Provost to Lead the Way,” The Chronicle of Higher Education, September 2012 This story from when Mitchell was first announced as vice provost contextualizes what the hire of a computer science professor for the position means for Stanford’s goals. “Vice Provost John Mitchell discusses online education initiatives,” The Stanford Daily, October 2012 In this interview with Stanford’s student newspaper, Mitchell addresses how Stanford’s approach to online learning compares to MIT and Harvard’s (he says Stanford’s is a faculty-driven process rather than top down) and responds to criticism that online instruction can’t compare to in-person classes.


“John Mitchell: How can we design for security?,” Stanford University School of Engineering, March 2017 Mitchell discusses current cybersecurity challenges: “The attacks that happen in the real world are the result of a tremendous amount of clever insight and ingenuity,” he says. “So, if we want to design something that's secure against the kind of attacks that a thoughtful, clever, creative, determined person would create, we can't really just do that by casual testing.” He says we need to test technology with the ways people will misuse it in mind. “Can Education Scale?” The University of Edinburgh, May 2016 Mitchell discusses the feasibility of low-cost education for all, which many people saw as an exciting opportunity after the emergence of Massive Open Online Courses (MOOCs). He considers how we should measure success in online courses as opposed to in-person ones and proposes the following test: “If someone is qualified to study the material at this level and devotes 1.5x effort, will they complete the course with 0.85x learning toward their future success?”

Keynote at the Future Computer Education Summit (FCES) sponsored by the China Computer Federation (CCF),