Microsoft Research's global academic research program providing Azure AI compute credits and access to cutting-edge foundation models for research advancing AI safety, human-AI interaction, and scientific discovery. The program operates through periodic Calls for Proposals, workshops, and conferences. Focus areas include AI safety and responsibility (robustness, transparency, evaluation methods), human-AI interaction (trust, creativity, productivity, reducing digital divides), and scientific discovery (knowledge discovery, hypothesis generation, multimodal data generation). The program is evolving into the Agentic AI Research and Innovation (AARI) initiative focusing on intelligent agent systems.

AI for Science & Safety Grant Program is a grant from the Foresight Institute that funds researchers, teams, and organizations using artificial intelligence as the primary driver of scientific breakthroughs or safety-focused innovation. The program provides grants of $10,000 to $100,000, along with compute resources and access to office and community space at new hubs in San Francisco and Berlin. It prioritizes mission-aligned teams pursuing ambitious, decentralized AI-driven progress across fields including biology, materials science, energy, and AI safety. Eligible applicants include academic researchers, independent teams, and organizations at any stage building AI-powered science or safety tools. The program is available to participants based near the San Francisco or Berlin hubs. Applications are reviewed on an ongoing basis; applicants should visit the Foresight Institute website for current submission details.

NIST AI Safety Research Grants is sponsored by National Institute of Standards and Technology (NIST). This program provides funding for research projects aimed at enhancing AI safety and reliability. Specific topics of interest include evaluating safe and trustworthy AI systems, and foundational research on examining, measuring, and understanding the tradeoff among various aspects of trustworthy AI systems, which include valid and reliable; safe, secure; and resilient; accountable; and transparent; explainable; and interpretable; privacy enhanced; and fair with security engineering and testing tools and metrics.