ETH Zürich (Swiss Federal Institute of Technology) is one of the world's leading technical universities, hosting several prominent AI safety and security research groups including the SPY Lab and SRI Lab.
ETH Zürich (Swiss Federal Institute of Technology) is one of the world's leading technical universities, hosting several prominent AI safety and security research groups including the SPY Lab and SRI Lab.
People
Updated 05/18/26Assistant Professor of Computer Science; head of SPY Lab
Full Professor of Computer Science; head of Secure, Reliable, and Intelligent Systems Lab
Funding Details
Updated 05/18/26- Annual Budget
- $2,000,000,000
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26ETH Zürich (Eidgenössische Technische Hochschule Zürich, or Swiss Federal Institute of Technology Zurich) is a public research university founded in 1855 by the Swiss Confederation. It is consistently ranked among the top universities globally in science and engineering, and 22 Nobel laureates have been affiliated with the institution. As of 2024, ETH Zürich enrolls approximately 26,200 students and employs over 13,600 staff across 16 departments. The university hosts several research groups directly relevant to AI safety and existential risk reduction. The Secure and Private AI (SPY) Lab, led by Assistant Professor Florian Tramèr, studies the worst-case behavior of deep learning systems from an adversarial perspective, aiming to understand and mitigate long-term threats to the safety and privacy of AI users. Tramèr completed his PhD at Stanford under Dan Boneh and a postdoc at Google Brain before joining ETH. The SPY Lab's AgentDojo benchmark won the CAIS SafeBench competition, and Tramèr has received a Google Research Scholar Award and an Amazon Research Award. The lab has around 11 members including PhD students and postdocs. The Secure, Reliable, and Intelligent Systems (SRI) Lab, led by Prof. Martin Vechev, runs the SafeAI project, which develops methods and tools for certifying and training robust, fair, and interpretable neural networks. The lab produced ERAN (ETH Robustness Analyzer for Neural Networks) and other widely-used tools, and spun out LatticeFlow, a commercial trustworthy AI platform. The SRI Lab publishes regularly at ICML, NeurIPS, ICLR, and other top venues. The ETH AI Center serves as a cross-departmental hub unifying AI research across more than 120 faculty professorships. It runs doctoral fellowship programs, hosts industry partnerships, and participates in the Swiss National AI Initiative (SNAI), which received CHF 20 million from the ETH Board for 2025-2028. ETH Zürich has been recognized as having the highest AI engineer density in Europe and hosts a growing community of AI safety researchers. ETH Zürich is state-funded by the Swiss Confederation, with total annual revenue of approximately CHF 2 billion (2024), of which CHF 1.4 billion comes from federal contributions and the remainder from third-party research funding, donations, and self-generated revenue.
Theory of Change
Updated 05/18/26ETH Zürich contributes to AI safety primarily through technical research: by studying adversarial attacks, robustness certification, and the security/privacy of ML systems, its researchers produce knowledge and tools that help developers build AI systems that are more reliable, resistant to manipulation, and auditable. The SPY Lab's adversarial approach — designing attacks to probe worst-case system behavior — generates findings that directly inform defenses and highlight risks before they are exploited. The SRI Lab's formal verification methods aim to provide mathematical guarantees about neural network behavior, a key ingredient for deploying AI safely in high-stakes settings. More broadly, ETH's role as a top-tier research institution means its publications, trained graduates, and open-source tools propagate into the wider AI development ecosystem, raising the baseline for safety practices globally.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26
ETH Zurich’s central hub for artificial intelligence, bringing together researchers on AI foundations, applications and implications across all departments to foster research excellence, industry innovation and AI entrepreneurship towards trustworthy, accessible and inclusive AI systems.
Research group in ETH Zurich’s Department of Computer Science that investigates the security, privacy and trustworthiness of machine learning systems, often designing adversarial attacks that probe worst‑case behaviour in order to ultimately improve the safety of these systems.
Research group in ETH Zurich’s Department of Computer Science focused on reliable, secure and trustworthy machine learning, with a particular emphasis on large language models and methods that make AI systems more robust and dependable.
Discussion
No comments yet. Be the first to share your thoughts.