Multilingual and Contextual AI Evaluation call for proposals - Fall 2026 supports researchers evaluating AI across languages and cultures to improve safety and fairness for underrepresented communities.
Funder: Amazon Research Awards
Due Dates (Anticipated): November 2027 (Submission deadline)
Funding Amounts: Cash and AWS Promotional Credits, each no more than $50,000 USD on average, for a one-year award period
Summary: Supports research on multilingual AI evaluation, cultural context, and emerging-market deployment, prioritizing underrepresented communities and the Global South.
Key Information: PIs must be full-time faculty at accredited institutions or permanent researchers at legally recognized nonprofit NGOs.
Amazon Research Awards (ARA) supports research that assesses and improves AI across languages and cultural contexts, particularly for underrepresented communities and the Global South. The call advances Amazon’s New Delhi Frontier AI Impact Commitments by expanding datasets, benchmarks, expertise, and evaluation infrastructure beyond high-resource languages. Responsible evaluation of safety and fairness, with human judgment involved, is central to the program.
The call welcomes low-resource language benchmarks, culturally contextual evaluation methods, and studies of AI deployment in emerging economies. Relevant work includes machine translation, question answering, summarization, conversational AI, localized alignment metrics, community feedback, and Indigenous knowledge. Related interests include scalable testing infrastructure, synthetic data and augmentation, cross-lingual transfer evaluation, and AI literacy and evaluation capacity building. Open-source datasets, benchmarks, and tools are a selection priority.