Arnav Arora

Arnav Arora is a PhD Fellow at the Department of Computer Science, University of Copenhagen (DIKU), specializing in Natural Language Processing. His work focuses on ethical AI, bias detection, and societal impacts of language models.

• Email: aar@di.ku.dk
• Location: Universitetsparken 1, 2100 København Ø

Arnav's research explores fine-grained value alignment in language models, harmful content detection, and cross-cultural differences in AI responses. His work bridges technical NLP advancements with social responsibility, including dual use ethical frameworks and community value analysis.

Key publication trends include:

• 2025: Bias mitigation through BiasGym framework
• 2024: Factcheck-Bench benchmark development
• 2023: Thorny Roses dual use analysis
• 2022: Cross-cultural value probing methods
• 2020: Multi-hop fact checking systems

Arnav contributes to the Software, Data, People & Society (SDPS) section, collaborating with interdisciplinary teams on projects involving language model evaluation and societal impact mitigation. His work often addresses real-world AI deployment challenges through academic-industry partnerships.