Center for AI Safety Research Grants

Center for AI Safety Research Grants is sponsored by Center for AI Safety. Provides grants for research on reducing societal-scale risks from AI, covering technical safety research, AI governance, and field-building initiatives.

Center for AI Safety Research Grants (Apply Now) | AI Safety Directory Center for AI Safety Research Grants Last updated : March 28, 2026 Funder Center for AI Safety The Center for AI Safety (CAIS) provides grants for research on reducing societal-scale risks from AI. Funding covers technical safety research, AI governance, and field-building initiatives.

CAIS has been instrumental in raising awareness of AI safety risks and supports researchers through direct grants, compute access, and collaborative research programs. alignment robustness governance responsible ai evaluation The Center for AI Safety (CAIS) has emerged as a central organization in the AI safety ecosystem, providing grants for research aimed at reducing societal-scale risks from advanced AI.

CAIS gained widespread attention with its 2023 statement on AI existential risk, signed by hundreds of leading AI researchers and public figures. The organization funds technical safety research, AI governance work, and field-building initiatives that strengthen the overall AI safety community. CAIS grants support a wide range of research activities, from formal alignment theory to practical safety benchmarking and evaluation.

The organization also operates compute access programs, providing researchers with the computational resources needed to conduct safety experiments on large models. CAIS's central position in the field means that funded researchers gain visibility and networking opportunities within a growing community of safety-focused scientists and policymakers.

CAIS accepts research proposals through its website, with both rolling submissions and periodic targeted calls for proposals on specific research topics. Applications should describe the research question, methodology, expected deliverables, and the team's qualifications.

CAIS is particularly interested in proposals that address concrete, measurable aspects of AI safety and that could produce results useful to the broader safety community. In addition to traditional research grants, CAIS offers compute grants providing access to GPU clusters for safety-critical experiments.

Researchers who need significant computational resources for their safety research can apply for compute access alongside or separately from research funding. CAIS also runs structured programs such as research workshops and collaborative research sprints that bring together researchers to work intensively on specific safety problems.

What Makes a Strong Application CAIS prioritizes proposals that address specific, well-defined risks from advanced AI systems and that propose rigorous methods for studying or mitigating those risks. Strong applications demonstrate a clear understanding of the current state of AI safety research and identify genuine gaps that the proposed work would fill.

Research with potential to produce public goods for the safety community, such as benchmarks, datasets, evaluation tools, or open-source safety libraries, is particularly valued. Applicants should have relevant technical expertise and a track record of producing high-quality research. However, CAIS also supports promising researchers who are newer to the field, especially through its more structured programs.

Proposals that could benefit from CAIS's compute resources should explain their computational requirements clearly and justify why large-scale experiments are necessary for the proposed research. Frequently Asked Questions Does CAIS provide compute access along with its research grants? Yes.

CAIS operates compute access programs that provide GPU clusters for AI safety research. Researchers can apply for compute access alongside their research grant application or separately. The program is designed to address the compute bottleneck that limits much safety research.

What is the typical funding range for CAIS grants? CAIS grants vary widely depending on the scope of the project. Smaller research projects may receive $25,000 to $100,000, while larger multi-year efforts can receive significantly more.

Compute grants are valued based on the GPU hours provided rather than direct monetary amounts. Can graduate students apply for CAIS grants? Yes.

CAIS supports researchers at all career stages, including graduate students. Student applicants should describe their advisor's involvement and any institutional support available for the project. CAIS also runs workshops and programs specifically designed for early-career researchers.

Open Philanthropy AI Safety Research Grants Major funder of AI safety research supporting alignment, governance, and technical safety work globally. Major private funder distributing $25M+ for AI safety research, governance, and policy work. AI safety company building reliable, interpretable AI systems and the Claude family of AI assistants.

AI research and deployment company working on safe and beneficial artificial general intelligence. Machine Intelligence Research Institute Nonprofit conducting foundational mathematical research on AI alignment and safety. Alignment Research Center Nonprofit researching AI alignment techniques, including eliciting latent knowledge and scalable oversight.

Open-source LLM vulnerability scanner that probes AI models for prompt injection, toxicity, and other weaknesses. Open-source text metrics toolkit for monitoring LLM application quality and safety in production. Microsoft's open-source CLI tool for security assessment of machine learning models.

Adversarial Robustness Toolbox IBM's comprehensive library for defending ML models against adversarial attacks across multiple frameworks. LLM Guardrails: The Complete Guide to AI Safety Guardrails (2026) Everything you need to know about LLM guardrails — what they are, why they matter, top tools, implementation patterns, and best practices for securing AI systems.

Prompt Injection Attacks: Types, Examples & Defenses A comprehensive guide to prompt injection attacks — how they work, the different types, real-world examples, and defense strategies for securing LLM applications. Blue Teaming in AI Security: Strategy, Tools & Best Practices A complete guide to AI blue teaming — the defensive operations function for monitoring, detecting, and responding to security threats against AI systems.

Get weekly AI security & safety updates New tools, frameworks, and research delivered to your inbox.