William Marino

My PhD supervisor is Nic Lane. Previously, I was co-supervised by Ross Anderson, who tragically passed in March 2024. Please take a moment to read about Ross's life and his titanic contributions to computer security and privacy here. 

My current research focus is on how we can build AI models and systems that can comply with AI regulation --- at scale and amid an increasingly complex AI ecosystem.

My research affiliations include the Machine Learning Systems Lab (CaMLSys), Security Group, and Leverhulme Centre for the Future of Intelligence (CFI). 

Prior to my PhD work, I practiced law and then worked in applied AI at Adobe and Stability AI. I was also co-founder and CEO of video AI startup Uru (acquired by Adobe). 

Outside of my research, I teach a graduate elective (5280: Productizing Machine Learning) at Cornell Tech and serve on the board of Kenya Connect, a non-profit that strives to advance education and health in rural Kenya. 

Trees, including fruit and nut trees, are a great passion of mine. In my home state of New Hampshire, I cultivate pawpaws (Asimina triloba). They've been called "America's best secret fruit." Have you ever tried one?

Select recent works

Papers
Marino, B., Hunter, R., Jamali, Z., Kalpakos, M.E., Kashyap, M., Hinton, I., Hanson, A., Nazir, M., Schnabl, C., Steffek, S., Wen, H., Lane, N.D. (2025). AIReg-Bench: Benchmarking Language Models That Assess AI Regulation Compliance. arXiv preprint arXiv:2510.01474.

Marino, B., Juels, A. (2025). Giving AI Agents Access to Cryptocurrency and Smart Contracts Creates New Vectors of AI Harm. arXiv preprint arXiv:2507.08249.

Schnabl, C., Hugenroth, D., Marino, B., Beresford, A.R. (2025.) Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments. arXiv preprint arXiv:2506.23706. (ICML-TAIG '25)

Yew, R.-J., Marino, B., Venkatasubramanian, S. (2025.) Red Teaming AI Policy: A Taxonomy of Avoision and the EU AI Act, FAccT '25. https://doi.org/10.1145/3715275.3732028. (FAccT '25)

Marino, B., Kurmanji, M., Lane, N.D. (2025.) Bridge the Gaps between Machine Unlearning and AI Regulation, arXiv preprint arXiv:2502.12430. (NeurIPS '25)

Sani, L., Iacob, A., Cao, Z., Marino, B., Gao, Y., Paulik, T., Zhao, W., Shen, W.F., Aleksandrov, P., Qiu, X., Lane, N.D. (2024.) Photon: Federated LLM Pre-Training. MLSys '25. (MLSys '25)

Iacob, A., Sani, L., Marino, B., Aleksandrov, P., Lane, N.D. (2024.), Worldwide Federated Training of Language Models, arXiv preprint arXiv:2405.14446. (NeurIPS-FL@FM '24)

Marino, B., Chaudhary, Y., Pi, Y., Yew, R.-J., Aleksandrov, P., Rahman, C., Shen, W.F., Robinson, I., Lane, N.D. (2024.) Compliance Cards: Automated EU AI Act Compliance Analyses amidst a Complex AI Supply Chain. arXiv preprint arXiv:2406.14758. (NeurIPS-RegML '24)

Corsi, G., Marino, B., Wong, W. (2024, June 3). The spread of synthetic media on X. Harvard Kennedy School Misinformation Review. 

Woisetschläger, H., Erben, A., Marino, B., Wang, S., Lane, N. D., Mayer, R., and Jacobsen, H.-A. (2024). Federated learning priorities under the European Union Artificial Intelligence Act. arXiv preprint arXiv:2402.05968. (ICML-GenLaw '24)

Blog posts

Marino, B. (2024, June 3). The EU AI Act’s Technical “Tension Areas.” Leverhulme Center for the Future of Intelligence Blog.

Corsi, G., Marconi, A., Marino, B. (2024, November 5). How Foundation Models Could Transform Synthetic Media Detection. Leverhulme Center for the Future of Intelligence Blog.

Talks

The EU AI Act - What It Means for You, in Simple Terms (Digital Tech Summit, October 2024)

Compliance Cards: Automated EU AI Act Compliance Analyses amidst a Complex AI Supply Chain (Stanford CodeX Weekly Meeting | September 12, 2024)

FL under AI Regulation (Flower AI Summit, March 2024)

Press

Kharif, O. Cornell Tech Professor Warns AI Agents And Crypto Spell Trouble (Bloomberg, July 29, 2025)

Gardner, R. Can We Use AI to Regulate AI? (University of Cambridge, Dept. of Computer Science and Technology, December 2024)