Alongside applied machine learning research, my recent work focuses on frontier AI governance, AI safety evaluation, and technical governance of advanced AI systems.
My interest in frontier AI governance is centered around mechanisms for tracking emerging capabilities and preparedness against risks, AI lifecycle assessment and evaluation, scalable oversight on AI R&D, public-private partnership and model release policies, and incident response. Selected works in progress include:
(i) frontier AI safety disclosure assessment, with the goal of comparing transparency, safety frameworks, and compliance; and
(ii) open source AI governance, risks from mis-developing AI, AI supply chain risks, and technical and behavioral recommendations for addressing risks.
I am also interested in standards-setting and best practices across the AI lifeycle — development and deployment. I served as the Secretary for IEEE Standards Association from 2024-2026 for standards-setting to measure energy usage of AI systems.
I presented a short paper on data governance encompassing the AI lifecycle at the AAAI 2025 Workshop on Datasets and Evaluators of AI Safety, and served on a workshop panel for safety considerations around agentic AI.
Peer-Reviewed Research Article on Data-centric practices in the AI/ Data lifecycle to improve Data and AI Governance:
- S. Chakraborty, “Data-Centric Safety and Ethical Measures for Data and AI Governance”, AAAI Workshop on Datasets and Evaluators of AI Safety, 2025, https://openreview.net/forum?id=iCcbtNIFuj
Conference Presentation and Preprint on the AI-Environment intersection and monitoring the multifaceted environmental impact of data centers
- Towards Monitoring Environmental Degradation of AI, Rethinking the Inevitability of AI Conference, Dec 2024
- S. Chakraborty, Towards A Comprehensive Assessment of AI’s Environmental Impact, 2024, In Review, preprint: https://arxiv.org/pdf/2405.14004
Other Policy-Relevant Activities: Public Comment submitted to the US AI Safety Institute and NIST on Safety Considerations for Chemical and/or Biological AI models in Dec 2024
Selected Policy Fellowships and Training:
From March-June 2026, I completed GovAI’s US AI Policy Program. I was a research associate at the Future Impact Group for Open Source AI Governance. I was selected for the yearlong training program in Science Policy under American Geophysical Union’s Voices for Science Program. I also participated in Aspen Digital’s Writing for Policy program in May 2024 and served as a research group member for Center for AI and Digital Policy from Sep-Dec 2024 and contributed to the AI Democratic Value Index Report, 2025.
Srija Chakraborty 