Our Research

Security, Privacy, and AI

As AI and other technologies advance, new vulnerabilities undermine user security, privacy, and trust. ICSI experts critically assess the security and privacy implications of new developments and advance solutions to protect users and systems.


Digital security concept

Smart solutions for trustworthy technology

Technology advancements bring great value to users and society at large, but they can also open new avenues for intrusion, surveillance, theft, and other risks. For decades, ICSI researchers have probed emerging technologies to identify and address critical vulnerabilities. Today, we are working at the forefront of AI safety and security in addition to addressing issues around privacy and trust for consumer applications such as medical apps and technologies used by children and teens.

Projects & Initiatives

Projects

Training Secure and Robust DNNs

We explore novel methods for training robust Deep Neural Networks (DNNs) that are free from security violations.

Projects

User Feedback for GenAI

We are developing a framework for providing user feedback on the outputs of generative AI features.

Projects

Bystander Privacy in Smart Homes

We study the privacy implications of smart home devices for bystanders—people who encounter these devices but are not their primary users.

Projects

AI Tools to Empower Adolescents

We address risks like cyberbullying, harassment, and predatory behavior with tools that are context-aware, participatory, and empowering.

Projects

Do Privacy Labels Work?

We are studying how privacy labels for mobile apps are applied and interpreted by developers and users.

Projects

Enhancing Security in LLMs

We build next-generation defenses to protect users and organizations from prompt-injection attacks on AI tools.

Projects

Software Privacy in Practice

We are studying how software developers understand and implement regulations aimed at protecting user privacy.

Projects

DNN Attack Detection

We are creating tools to detect backdoor attacks on Deep Neural Networks (DNNs).

Projects

Privacy in Mobile Health

We study how users perceive privacy regulations and protections and how developers implement those protections in mobile health apps.

Projects

Informing Accountable Software

We are applying agent-based modeling to inform the design and regulation of software systems that are accountable to the legal and social contexts in which they are implemented.

Projects

Co-Designed Online Safety for Teens

We advance online safety for teens by conducting human-centered, interdisciplinary research and developing strategies to build more resilient digital ecosystems.

Projects

Detecting Adolescent Online Risks

We develop privacy-preserving, context-aware algorithms to improve the accuracy and fairness of online risk detection.

Latest News

Our Experts

Our experts conduct visionary research, mentor and collaborate with our fellows and visitors, and are frequently invited to serve as keynote speakers and panelists. In addition, ICSI experts are available for technical and strategy consulting, allowing you to apply ICSI brainpower to your organization’s most pressing challenges across many areas of digital security and privacy.

Join ICSI

ICSI offers the best of all worlds—a place to do transformative work without unnecessary bureaucracy. With no teaching obligations, minimal red tape, and flexible employment models, you’re free to do your best work, your way. 

Visit ICSI

ICSI is a global research community that feels like home. A visit to ICSI puts you at the center of the action to collaborate with top researchers in a world-class innovation ecosystem. 

Get in touch

Want to discuss opportunities to work with ICSI? We’d love to hear from you.

2150 Shattuck Ave., #250
Berkeley, CA 94704

+1 (510) 666-2900

contact @ icsi.berkeley.edu