PL+Law

Can we infer entailments from legal documents and test whether people understand the implications of what they agree to?

Backstory!

In 2021, I began working with Dorsa Mohammadi Arezooji on a new domain in the nascent PL+Law field. We were inspired by Basu et al.'s Property Conveyances as Programming Languages. We began by investigating two domains: arbitration agreements in EULAs and insurance contracts. We were seeking out domain experts in the legal field and have outlined projects in NLP and user comprehension of privacy concerns. Unfortunately due to factors outside her control, Dorsa was not able to pursue her PhD with me.

Erin McBride, now a 1L at University of Wisconsin Law School practicing lawyer(!), worked with me in this space, researching contracts of adhesion and their role in changes to websites under experimentation conditions.

Current work

I had recently been able to pursue this work under the supervision of Dr. Chris Martens at Northeastern University, as part of Dr. Martens' CAREER award. While our research questions were initially formal in nature, to ground them in real-world scenarios, we took a bit of a detour to better understand how Fediverse administrators who are not necessarily compliance officers understand documents such as privacy policies. Several students have worked on and continue to work on this topic.

Future work

Dr. Martens and I remain interested in the question of what formal methods can provide to Fediverse administrators as end-users. We are particularly interested in collaborating with experts in policy and law, including both experts in regulatory compliance (for when we treat administrators as e.g. fiduciaries) and common law (for when we treat administrators as community members involved in disputes). If you'd like to discuss the current and/or future status of this project, please reach out.