Amazon Research Awards Program (ARAP)

Overview

In February 2026, Amazon signed the New Delhi Frontier AI Impact Commitments at the India AI Impact Summit. The second commitment, "Strengthening Multilingual and Use-Case Evaluations," asks signatories to:

• Collaborate with governments and local ecosystems to develop datasets, benchmarks, and expertise for evaluating AI systems in underrepresented languages and cultural contexts
• Prioritize the Global South in this work
• Support evaluation infrastructure that does not yet exist at scale

This ARA Call for Proposals is how we're making progress on that commitment. We're looking for proposals from academic institutions and non-profits that tackle multilingual AI evaluation, culturally contextual AI, or AI deployment in emerging markets.

Amazon already invests in multilingual and cross-lingual AI through efforts like Bedrock Cross-Lingual Evaluation, SWE-PolyBench, Amazon Translate, and AWS-enabled projects like Kiwa. This call extends that work by funding external research, with a particular focus on outreach to include institutions in the Global South.

Areas of interest: We're looking for proposals in the following areas, especially work that benefits underrepresented languages, cultures, and regions:

• Benchmarks for under-represented languages
• Culturally contextual AI evaluation
• Application of AI in emerging markets
• Additional welcome topics
• Other related topics are also welcome, including but not limited to:
  • Low-resource data collection methods, including synthetic data generation and data augmentation for underrepresented languages
  • Human-in-the-loop and community-sourced evaluation
  • Responsible AI for multilingual systems - fairness, safety, and transparency across languages
  • Cross-lingual transfer learning and its evaluation
  • AI literacy and capacity building for evaluation in Global South institutions

Selection criteria: ARA will make funding decisions based on the following criteria:

• Potential impact on multilingual AI evaluation, particularly for underrepresented languages and the Global South
• Scientific rigor, novelty, and quality of the proposed methodology. Rigor of the responsible-AI approach in the evaluation of multilingual and culturally contextual AI systems
• Feasibility of the research plan within the one-year award period
• Commitment to open-source release of datasets, benchmarks, and tools
• Alignment with the New Delhi Frontier AI Impact Commitments & partnership with institutions, governments, or local communities.
• Use of AWS AI/ML services (e.g., Amazon SageMaker, Amazon AI services, Amazon Bedrock)

Each Principal Investigator is permitted to submit only one proposal to the ARA Program per call for proposal period.

Timeline
Submission period: October 1 - November 4, 2026 (11:59PM Pacific Time).
Decision letters will be sent out in February 2027.

Award details
Selected Principal Investigators (PIs) may receive the following:

• Unrestricted funds, no more than $50,000 USD on average
• AWS Promotional Credits, no more than $50,000 USD on average
• Training resources, including AWS tutorials and hands-on sessions with Amazon scientists and engineers

Awards are structured as one-time unrestricted gifts. The budget should include a list of expected costs specified in USD, and should not include administrative overhead costs. The final award amount will be determined by the awards panel.

All Award amounts will be determined by Amazon in its sole discretion. Any cash component of an Award:

1. will be structured as a one-time unrestricted gift to your Principal Investigator’s academic institution or organization;
2. will be provided directly to your academic institution or organization for distribution and management; and

may not be used for indirect expenses which are not allocable, reasonable, adequately documented, and consistent with established policies and practices of your academic institution or organization.