Surrey Workshops on Jurisprudence and AI: First Edition
Wednesday 18 November 2026, 14:00–17:00 GMT (09:00–12:00 US Eastern), online. Hosted jointly by the Surrey Centre for Law and Philosophy and the Surrey Law and Technology Hub, University of Surrey.
The Surrey Workshops on Jurisprudence and AI are a new series of online workshops on the conceptual and normative questions that artificial intelligence raises for law. Each workshop discusses work in progress in depth. Papers are circulated in advance and taken as read, so there are no presentations. A designated commentator opens each session, followed by general discussion.
To take part, request a place on Eventbrite. Registration is a request, not a confirmation: the organisers will confirm places by email, and confirmed participants will receive the papers and the joining link about a week in advance. Participants are expected to have read the papers.
Programme
14:00–14:05 Welcome
14:05–15:20 Session 1
Mihailis Diamantis (University of Iowa) and Sebastian Benthall (NYU School of Law), “AI Wardens”
In a very probable, not-too-distant future, humans will use artificial intelligence to carry out essential economic and social exchange. Such AI tools are often called “AI agents” because they extend our capacities, allowing us to do more, faster. Well-meaning scholars have argued that developers should limit the potential harms of AI by designing agents that strictly adhere to law. However, as society’s use of AI agents leads to reliance on them, the same tools that extend our agency will also become limits on it. AI agents that will not break the law become “AI wardens” that foreclose our capacity to do the same. This Essay advocates for the moral and political imperative of allowing AI to break the law at the request of its users. At stake is a fundamental check on state abuse: the power to disobey. Users should be able to compel their AI agents to break the law, but only after explicitly accepting responsibility for the ensuing violation. Without this space to refuse the state’s command, law-following AI wardens risk tipping society into effortless totalitarianism.
15:20–15:30 Break
15:30–16:45 Session 2
Tianqi Shi and Mark Dsouza (UCL), “Offences against the (online) person”
Our thesis is that there is a pro tanto case for criminalising certain wrongs done to a person’s avatar, or to their curated online persona or profile, as offences against the person. To that end, we argue first, that especially given technological advances, a corporeal person’s relationship with their avatar can go beyond the relationship an owner has with their property, such that the avatar becomes an extension of the person’s agency. Thus, wrongfully harming that ‘extended person’ can also be an offence against the person. And second, we argue that the online world is a valuable common resource, and access to it deserves the state’s protection. Since access to much of it requires an avatar, a person’s reasonably unfettered access to their avatar also deserves protection. One way of offering that protection is to reject the notion that an unwillingness to give up one’s avatar, and with it the online world, is legally valid acceptance of the risk of harm.
16:45–17:00 Close