Ideas: More AI-resilient biosecurity with the Paraphrase Project @MicrosoftResearch
Ideas: More AI-resilient biosecurity with the Paraphrase Project  @MicrosoftResearch
Uploaded October 2025 | Updated November 2026, 2 hours ago
Microsoft’s Eric Horvitz and guests Tessa Alexanian, James Diggans, and Bruce Wittmann discuss the Paraphrase Project—a red-teaming effort that exposed and secured a biosecurity vulnerability in AI-driven protein design. The work offers a model for addressing AI’s dual-use risks.

Show notes: microsoft.com/en-us/research/podcast/ideas-more-ai-resilient-biosecurity-with-the-paraphrase-project
Ideas: More AI-resilient biosecurity with the Paraphrase ProjectuGAP: A Unified Geospatial AI Platform for African Geographic DataVoluMe: Authentic 3D Video Calls from Live Gaussian Splat PredictionWhat AIs impact on individuals means for the health workforce and industryAI Frontiers: Rethinking intelligence with Ashley Llorens and Ida MomennejadTruthTable: A Verifiable Query EngineDetecting and Mitigating Bias in Voice Activated TechnologiesBeyond Swahili: Designing Inclusive AI for Bantu LanguagesAbstracts: NeurIPS 2024 with Pranjal ChitaleScaling Density Functional Theory with Gaussian SplattingTool-space Interference: An emerging problem for LLM agentsEpisode 7: The road ahead
Microsoft Research |

Ideas: More AI-resilient biosecurity with the Paraphrase Project

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER