Security, privacy, and cryptography
Career Opportunity
Member of Technical Staff, AI Safety Post-Training – MAI Super Intelligence Team
As a Member of Technical Staff, AI Safety Post-Training, you will work to develop and implement cutting-edge safety methodologies for post-training large language models with agentic and reasoning capabilities that are served to millions of…
Video
Publicly-verifiable elections
Microsoft’s free, open-source ElectionGuard tools enable voters to verify their votes were accurately counted without compromising privacy or trusting election equipment or personnel. New research eliminates the need for cryptographic keys, making the process far…
Career Opportunity
Research Intern – Cryptography
The researchers and engineers in the Cryptography team pursue challenging research that has an impact at Microsoft and the world at large. Most recently we have focused on cryptographic identity, formally verified cryptography, encrypted communications,…
Microsoft Research Blog
Reducing Privacy leaks in AI: Two approaches to contextual integrity
New research explores two ways to give AI agents stronger privacy safeguards grounded in contextual integrity. One adds lightweight, inference-time checks; the other builds contextual awareness directly into models through reasoning and RL.
Publication
CHERI-Lite for Memory Safety Exploit Mitigation
Video
Designing safe digital systems for the humanitarian sector
Speaker: Carmela Troncoso Host: Betül Durak In this talk we overview our collaboration with the International Committee of the Red Cross, in which we help them to digitalize their aid distribution process without increasing risks…
Microsoft Research Blog
BlueCodeAgent: A blue teaming agent enabled by automated red teaming for CodeGen AI
BlueCodeAgent is an end-to-end blue-teaming framework built to boost code security using automated red-teaming processes, data, and safety rules to guide LLMs’ defensive decisions. Dynamic testing reduces false positives in vulnerability detection.
